Upload
others
View
29
Download
5
Embed Size (px)
Citation preview
Antenna Positioning Analysis and
Dual-Frequency Antenna Design of
High Frequency Ratio for Advanced
Electronic Code Responding Labels
by
Kin Seong Leong
B.E. (Electrical & Electronic, First Class Honours),The University of Adelaide, Australia, 2003
Thesis submitted for the degree of
Doctor of Philosophy
in
The School of Electrical & Electronic Engineering,
The Faculty of Engineering, Computer and Mathematical Sciences,
The University of Adelaide, Australia
January 2008
c© 2008
Kin Seong Leong
All Rights Reserved
To
My Parents
Page iv
Contents
Contents v
Abstract xv
Statement of Originality xvii
Acknowledgment xix
Conventions xxi
Publications xxiii
List of Figures xxvii
List of Tables xxxvii
Chapter 1. Thesis Introduction 1
1.1 Thesis Title . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Background of Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Thesis Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6 Chapter Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Chapter 2. RFID Systems 11
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 History of RFID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 A Simple RFID System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Page v
Contents
2.4 Variants of RFID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.1 Active and Passive RFID Tags . . . . . . . . . . . . . . . . . . . . . 14
2.4.2 Backscattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.3 Common Frequency Bands for RFID Operations . . . . . . . . . . 16
2.5 RFID and Supply Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.6 The EPC Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6.1 The Early Stage EPC Network . . . . . . . . . . . . . . . . . . . . 19
2.6.2 The EPC Network Current State . . . . . . . . . . . . . . . . . . . 21
2.7 RFID Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Chapter 3. Path Loss and Position Analysis 27
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2 Path Loss Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2.1 Free Space Path Loss . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2.2 In-building Path Loss . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4 Path Loss Model for RFID . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.5 Simple Two Reader Interference . . . . . . . . . . . . . . . . . . . . . . . . 34
3.6 Background for RFID Positioning Analysis . . . . . . . . . . . . . . . . . 37
3.6.1 Power Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.6.2 Antenna Gain Pattern . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.6.3 Frequency Spectrum Channelling . . . . . . . . . . . . . . . . . . 40
3.6.4 Listen Before Talk . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.7 Simulation Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.7.1 Simulation Concepts Example . . . . . . . . . . . . . . . . . . . . 47
3.8 Results and Discussion (RITR) . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.8.1 Modification of Existing Software . . . . . . . . . . . . . . . . . . 48
Page vi
Contents
3.8.2 Guidelines to Avoid RITR . . . . . . . . . . . . . . . . . . . . . . . 50
3.9 Results and Discussion (RRS) . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.9.1 One Antenna Simulation . . . . . . . . . . . . . . . . . . . . . . . 53
3.9.2 Two Antenna Simulation . . . . . . . . . . . . . . . . . . . . . . . 55
3.9.3 One Antenna in Horizontal Position . . . . . . . . . . . . . . . . . 56
3.9.4 Antennas Operating in Different Channels . . . . . . . . . . . . . 57
3.9.5 A More Hostile Environment . . . . . . . . . . . . . . . . . . . . . 63
3.10 Real Life Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.10.1 A Checkout Counter . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.10.2 Reader Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . 64
3.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Chapter 4. RFID Operational Considerations 67
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.2 RFID EMC Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.2.1 Frequency Hopping Spread Spectrum (FHSS) . . . . . . . . . . . 70
4.2.2 Listen Before Talk (LBT) . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3 RFID Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.4 Path Loss Measurement Experiment . . . . . . . . . . . . . . . . . . . . . 71
4.4.1 Experimental Setup: . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.4.2 Room Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.4.3 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.5 Sources of Simulation Error . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.5.1 Path Loss Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.5.2 Reflection, Refraction, and Diffraction . . . . . . . . . . . . . . . . 82
4.5.3 Radiation Pattern of Antenna . . . . . . . . . . . . . . . . . . . . . 84
Page vii
Contents
4.5.4 Simulation Result Interpretation and Analysis . . . . . . . . . . . 85
4.6 Second Carrier Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.6.1 Effect of the Position of Tags . . . . . . . . . . . . . . . . . . . . . 90
4.6.2 Effect of Environment Factor . . . . . . . . . . . . . . . . . . . . . 91
4.6.3 Combining First and Second Carrier Sensing . . . . . . . . . . . . 92
4.7 Investigation of Specific Absorption Rate (SAR) . . . . . . . . . . . . . . 96
4.7.1 SAR Level for UHF RFID Operation . . . . . . . . . . . . . . . . . 96
4.7.2 Experiment Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.7.3 Simulation of Power Density . . . . . . . . . . . . . . . . . . . . . 97
4.7.4 Maximum Power for Every Channel . . . . . . . . . . . . . . . . . 98
4.7.5 Worst-Case Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.7.6 Recommendations on SAR . . . . . . . . . . . . . . . . . . . . . . 102
4.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Chapter 5. Reader Synchronisation 105
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.2.1 ETSI 302 208 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.2.2 EPC Class 1 Generation 2 Protocol . . . . . . . . . . . . . . . . . . 107
5.2.3 Problem in Dense Reader Environment . . . . . . . . . . . . . . . 108
5.3 Reader Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.4 Actual Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.4.1 Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.4.2 Positioning of LBT Sensors . . . . . . . . . . . . . . . . . . . . . . 111
5.4.3 Antenna Positioning . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.5 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.6 Synchronisation Fine-tuning . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Page viii
Contents
5.6.1 Reduction of Output Power . . . . . . . . . . . . . . . . . . . . . . 116
5.6.2 Reduction of Overall Reader Talking Time . . . . . . . . . . . . . 117
5.6.3 Use of External Sensors . . . . . . . . . . . . . . . . . . . . . . . . 117
5.6.4 RF Opaque or RF Absorbing Materials . . . . . . . . . . . . . . . . 118
5.6.5 Frequent Rearrangement of Channels . . . . . . . . . . . . . . . . 118
5.7 Variation of Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.7.1 Separation of Transmitting and Receiving Channels . . . . . . . . 119
5.7.2 Separation of RFID and Non-RFID Signals . . . . . . . . . . . . . 121
5.8 Updated Progress on Development of RFID Reader Synchronisation . . 121
5.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Chapter 6. RFID Tag Antenna Design 123
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.2 Antenna Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.3 Antenna Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
6.4 Resonant Circuit for Antenna . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.5 Challenges in RFID Tag Antenna Design . . . . . . . . . . . . . . . . . . . 140
6.6 RFID Chips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6.7 RFID Readers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.7.1 HF Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.7.2 UHF Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.8 RFID Tag Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.9 Antenna Design and Simulation . . . . . . . . . . . . . . . . . . . . . . . 150
6.9.1 Ansoft HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.9.2 Scripting in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.9.3 Plotting in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.9.4 ISO-Pro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Page ix
Contents
6.10 HF RFID Antenna Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
6.10.1 A Simple Loop Antenna . . . . . . . . . . . . . . . . . . . . . . . . 161
6.10.2 A HF Planar Spiral Coil Antenna . . . . . . . . . . . . . . . . . . . 162
6.10.3 An HF Antenna for a New Wine Closures . . . . . . . . . . . . . . 167
6.10.4 An HF Antenna for a Pigs . . . . . . . . . . . . . . . . . . . . . . . 168
6.11 UHF RFID Antenna Design . . . . . . . . . . . . . . . . . . . . . . . . . . 179
6.11.1 A Common UHF RFID Tag . . . . . . . . . . . . . . . . . . . . . . 181
6.11.2 Simple Planar Dipole . . . . . . . . . . . . . . . . . . . . . . . . . . 182
6.11.3 A UHF Antenna for Sheep . . . . . . . . . . . . . . . . . . . . . . . 183
6.11.4 A UHF Antenna for Pigs . . . . . . . . . . . . . . . . . . . . . . . . 185
6.11.5 A UHF Antenna for Beer Kegs . . . . . . . . . . . . . . . . . . . . 187
6.11.6 A UHF Antenna for New Wine Closures . . . . . . . . . . . . . . 188
6.12 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Chapter 7. Dual-Frequency RFID Antenna 191
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
7.2 Current Dual-Frequency Antenna Design . . . . . . . . . . . . . . . . . . 193
7.3 Design Aims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
7.4 Dual-Frequency Antenna Design . . . . . . . . . . . . . . . . . . . . . . . 196
7.4.1 Independent HF and UHF Antenna Design . . . . . . . . . . . . . 196
7.4.2 Quick Feasibility Test . . . . . . . . . . . . . . . . . . . . . . . . . . 198
7.4.3 Tunability Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
7.4.4 Redesign of the UHF Dipole . . . . . . . . . . . . . . . . . . . . . 206
7.4.5 Compatibility of Dipole in HF . . . . . . . . . . . . . . . . . . . . 210
7.4.6 Merging of the New UHF Dipole with the HF Coil . . . . . . . . . 212
7.5 Antenna Fabrication and Testing . . . . . . . . . . . . . . . . . . . . . . . 217
7.6 Miniaturisation of Dual-Frequency RFID Antenna with High Frequency
Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Page x
Contents
7.7 Improvement on Novel Dual-Frequency Antenna . . . . . . . . . . . . . 221
7.8 Testing on Miniaturised Antenna . . . . . . . . . . . . . . . . . . . . . . . 230
7.9 Final Design of Miniaturised Dual-Frequency Antenna . . . . . . . . . . 231
7.10 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Chapter 8. Alternative Dual-frequency Antenna Designs 235
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
8.2 Alternate Methods in Merging an HF Antenna and a UHF Antenna . . . 236
8.2.1 Different Feed Point Location . . . . . . . . . . . . . . . . . . . . . 237
8.2.2 Different UHF Antenna Types . . . . . . . . . . . . . . . . . . . . 239
8.3 HF Antenna Acting as UHF Antenna at UHF . . . . . . . . . . . . . . . . 241
8.3.1 Loss in Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
8.3.2 Antenna Measurement Results . . . . . . . . . . . . . . . . . . . . 245
8.4 UHF Antenna Acting as an HF Antenna at HF . . . . . . . . . . . . . . . 246
8.5 Multi-Feed Point Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
8.5.1 Stacking UHF and HF Antennas . . . . . . . . . . . . . . . . . . . 247
8.5.2 Side-by-side HF and UHF Antenna . . . . . . . . . . . . . . . . . 251
8.6 Comparison Between Different Methodologies . . . . . . . . . . . . . . . 251
8.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Chapter 9. Measurement Techniques 255
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
9.2 Background and Literature Review . . . . . . . . . . . . . . . . . . . . . . 257
9.3 Settings and Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
9.4 Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
9.4.1 A Balanced Bow Tie Antenna . . . . . . . . . . . . . . . . . . . . . 262
9.4.2 Half Bow Tie Antenna on Ground Plane . . . . . . . . . . . . . . . 264
9.4.3 Results Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . 264
Page xi
Contents
9.5 Discussion and Improvement . . . . . . . . . . . . . . . . . . . . . . . . . 265
9.6 Effect of Inaccuracy in Measurement . . . . . . . . . . . . . . . . . . . . . 268
9.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Chapter 10.Thesis Conclusions 273
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
10.2 Review of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
10.3 Possible Research Extensions . . . . . . . . . . . . . . . . . . . . . . . . . 275
10.4 Summary of Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
10.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
Appendix A. Calculation of Impedance of Coplanar Strip Line 279
Appendix B. Some Notes on the Concept of Inductance 281
Appendix C. MATLAB Code for Path Loss Calculation 285
C.1 Main Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
C.2 Antenna Gain Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
C.3 Path Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Appendix D. MATLAB Code for HFSS VB Script Generation 291
Appendix E. MATLAB Code for Inductance Calculation 297
E.1 Main Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
E.2 Inductance Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
E.3 Mutual Inductance - Positive . . . . . . . . . . . . . . . . . . . . . . . . . 302
E.4 Mutual Inductance - Negative . . . . . . . . . . . . . . . . . . . . . . . . . 303
Appendix F. Path Loss Experiment 305
F.1 Preliminary Setting Up Procedure . . . . . . . . . . . . . . . . . . . . . . 305
F.2 Experiment Procedure and Results . . . . . . . . . . . . . . . . . . . . . . 306
Page xii
Contents
F.2.1 Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
F.2.2 Signal Strength at Different Distances (Within and Between Build-
ings) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
F.2.3 Reflection from Typical Wall and a Conductive Fence . . . . . . . 310
F.2.4 Propagation Loss Outdoors . . . . . . . . . . . . . . . . . . . . . . 312
F.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Appendix G. RFID Deployment in Piggery 315
G.1 Modelling of Pig Feeder . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
G.2 Final Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
G.2.1 Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . 317
G.2.2 Design of RFID Tag Antennas . . . . . . . . . . . . . . . . . . . . . 318
G.2.3 Feeder Design and Experiment Setup . . . . . . . . . . . . . . . . 322
G.2.4 Network Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
G.2.5 Experiment Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
G.2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Bibliography 333
Page xiii
Page xiv
Abstract
The research background of this thesis is Radio Frequency Identification (RFID), where
an object can be identified remotely using electromagnetic waves. The focus of this the-
sis is on the in-depth investigation of two major problems in the RFID deployment in
supply chain applications, namely the reader collision problem in dense reader envi-
ronments and the tag performance problem in hostile environments.
To resolve the reader collision problem, the first part of this thesis offers a compre-
hensive path loss model for the analysis of the positioning of RFID reader antennas.
Simulation software was developed to predict the signal strength at a certain distance
from a reader antenna in a dense reader environment.
This simulation software was also utilised to publish insights and research results in
four major areas, which are: (i) Investigation on the sources of error in RFID simula-
tion, to provide sensible and meaningful simulation results before actual deployment
of RFID readers. (ii) The development of the idea of reader synchronisation, mainly
to address the strict regulations imposed on the deployment of RFID readers in Eu-
rope. (iii) The determination of the threshold value for second carrier sensing in RFID,
to enable the proper enforcement of second carrier sensing to avoid tag confusion in
dense reader environments. (iv) The examination of Specific Absorption Rate (SAR) to
ensure human safety in a dense RFID reader environment.
The second part of this thesis addresses the RFID tag performance problem in hostile
environments. The focus is on the development of HF and UHF tags, from the initial
tag antenna design, tag antenna simulation, tag antenna prototyping and measure-
ment, to the manufacturing of fully functional RFID tags at laboratory standards by
combining RFID chips on to tag antennas.
Page xv
Abstract
Though there are existing commercial grade HF and UHF RFID tags, they are mostly
aimed at pallet level applications and are not suitable for deployment in hostile en-
vironments. The study cases presented in this thesis are mostly industrially driven,
where there is a need to design specialty HF and UHF tag antennas.
With a strong foundation in the development of HF and UHF RFID tags for various
industrially driven applications, the research then concentrates on the development
of a novel dual-frequency RFID antenna, which operates in both the HF and UHF
regions. This dual-frequency RFID tag antenna embraces the benefits of both the HF
and UHF tag antenna, which enable it to have a good read range while operating in
environments that pose difficulties for RFID technology, for example applications in
which ionised liquid is present, such as in cases of wine or bottled drinks.
Several methodologies were used to develop a dual-frequency antenna, including the
merging of HF and UHF antennas, and having a UHF resonance point on a typical HF
antenna. With the successful development of an original dual-frequency antenna, the
research was then expanded to miniaturise this dual-frequency antenna.
The benefits of RFID deployment in supply chains are undoubtedly massive, though
there are still issues and challenges to be resolved before a world-wide adoption is
possible. This thesis contributes in recommending various reader antenna positioning
and deployment techniques, and also contributes in developing HF tag antennas and
UHF tag antennas for hostile environments, and a novel dual-frequency tag antenna
to progress towards the aim of ubiquitous object identification.
Page xvi
Statement of Originality
This work contains no material that has been accepted for the award of any other de-
gree or diploma in any university or other tertiary institution and, to the best of my
knowledge and belief, contains no material previously published or written by an-
other person, except where due reference has been made in the text.
I give consent to this copy of my thesis being made available in the University Li-
brary.
The author acknowledges that copyright of published works contained within this
thesis (as listed in the publications page) resides with the copyright holder/s of those
works.
Signed Date
Page xvii
Page xviii
Acknowledgment
I would like to thank my principal supervisor Prof. Peter Cole for his guidance through-
out my research candidature. His constant encouragement and support have made the
completion of this thesis possible.
Also, I would like to acknowledge Prof. Bevan Bates and Dr. Samuel Mickan (my co-
supervisors) for their countless advice and assistance, and Alfio Grasso for his help in
arranging projects with industrial partners.
Many thanks to the administrative and technical support staff of School of Electrical
& Electronic Engineering, University of Adelaide, especially to Geoff Pook who fabri-
cated most of my prototype antennas.
Sincere thanks to anyone who has contributed in any technical discussion with me,
notably my fellow colleagues from Auto-ID Laboratory, and David Hall.
I am very grateful for my postgraduate studies financial supports, an Australian Post-
graduate Awards (APA) and the Frank Perry Scholarship. Also, special thanks to Pork
CRC for funding the RFID development and field trial in a piggery.
Also, I wish to extend my gratitude to industry partners who offered their help and
support in completing some of my projects, especially Dr. Ian McCauley from DPI
Victoria, and Bruce Dumbrell from Leader Products.
Last but not least, I really appreciate the unconditional love and support from my
family, and also from Mun Leng, my research partner and fiancee.
Kin Seong Leong
Dec. 2007
Page xix
Page xx
Conventions
Typesetting
This thesis is typeset in Times New Roman and Sans-Serif using LATEX2e. Referenc-
ing and citation style are based on the Institute of Electrical and Electronics Engineers
(IEEE) Transaction style [1]. For an electronic source, the last updated date of the source
is enclosed within round parentheses, and is placed immediately behind the author(s)’
name [2]. The last access date is included within square parentheses and can be found
at the end of the entry.
Units
The International System of Units (abbreviated SI units) [3] is used in this thesis. Pre-
fixes “nano”, “micro”, and “milli” are preferred but prefix “cm” is avoided.
Spelling
English spelling in this thesis is based on Australian English. One exception is in some
special cases where the proper noun is used. For example: “Auto-ID Center”, not
“Auto-ID Centre”. Also, the plural of “antenna” is chosen to be “antennas” not “an-
tennae”, to be in line with most of the international technical publications.
Page xxi
Page xxii
Publications
Book Chapter
[1] K. S. Leong, M. L. Ng, and P. H. Cole, “RFID reader synchronisation,” in RFID Handbook: Applica-
tions, Technology, Security, and Privacy, S. Ahson and M. Ilyas, Eds. CRC, 2008.
[2] M. L. Ng, K. S. Leong, and P. H. Cole, “RFID tags for metallic object identification,” in RFID Hand-
book: Applications, Technology, Security, and Privacy, S. Ahson and M. Ilyas, Eds. CRC, 2008.
Journal
[1] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Dense RFID reader deployment in Europe using
synchronization,” Journal of Communications, vol. 1, no. 7, pp. 9–16, 2006.
Conference
[1] K. S. Leong, M. L. Ng, and P. H. Cole, “HF and UHF RFID tag design for pig tagging,” in 11th
Biennial Conference of the Australasian Pig Science Association (APSA), Brisbane, Australia, 25-28 Nov.
2007.
[2] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation on the deployment of HF and UHF RFID
tag in livestock identification,” in IEEE Antennas and Propagation Society International Symposium,
Honolulu, Hawaii, USA, 10-15 Jun. 2007.
[3] K. S. Leong, M. L. Ng, and P. H. Cole, “Miniaturization of dual-frequency RFID antenna with
high frequency ratio,” in IEEE Antennas and Propagation Society International Symposium, Honolulu,
Hawaii, USA, 10-15 Jun. 2007.
[4] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of RF cable effect on RFID tag antenna
impedance measurement,” in IEEE Antennas and Propagation Society International Symposium, Hon-
olulu, Hawaii, USA, 10-15 Jun. 2007.
[5] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of the threshold of second carrier sensing in
RFID deployment,” in 2006 International Symposium on Applications and the Internet (SAINT) Work-
shop, RFID and Extended Network: Deployment of Technologies and Applications, Hiroshima, Japan,
15-19 Jan. 2007.
Page xxiii
Publications
[6] K. S. Leong, M. L. Ng, and P. H. Cole, “Dual-frequency antenna design for RFID application,” in
21st International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC
2006), Chiang Mai, Thailand, 10-13 July 2006.
[7] K. S. Leong, M. L. Ng, and P. H. Cole, “Operational considerations in simulation and deployment of
RFID systems,” in 17th International Zurich Symposium on Electromagnetic Compatibility, Singapore,
27 Feb. - 3 Mar. 2006.
[8] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Synchronisation of RFID readers for dense
RFID reader environments,” in 2006 International Symposium on Applications and the Internet (SAINT)
Workshop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Arizona,
USA, 23-27 Jan. 2006.
[9] K. S. Leong, M. L. Ng, and P. H. Cole, “Positioning analysis of multiple antennas in a dense RFID
reader environment,” in 2006 International Symposium on Applications and the Internet (SAINT) Work-
shop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Arizona,
USA, 23-27 Jan. 2006.
[10] K. S. Leong, M. L. Ng, and P. H. Cole, “The reader collision problem in RFID systems,” in IEEE
2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless
Communications, Beijing, China, 8-12 Aug. 2005.
[11] M. L. Ng, K. S. Leong, and P. H. Cole, “Design and miniaturization of an RFID tag using a sim-
ple rectangular patch antenna for metallic object identification,” in IEEE Antennas and Propagation
Society International Symposium, Honolulu, Hawaii, USA, 10-15 Jun. 2007.
[12] M. L. Ng, K. S. Leong, and P. H. Cole, “A small passive UHF RFID tag for metallic item identifi-
cation,” in 21st International Technical Conference on Circuits/Systems, Computers and Communications
(ITC-CSCC 2006), Chiang Mai, Thailand, 10-13 July 2006.
[13] M. L. Ng, K. S. Leong, D. M. Hall, and P. H. Cole, “A small passive UHF RFID tag for livestock
identification,” in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC
Technologies for Wireless Communications, Beijing, China, 8-12 Aug. 2005.
[14] M. L. Ng, K. S. Leong, and P. H. Cole, “Analysis of constraints in small UHF RFID tag design,”
in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for
Wireless Communications, Beijing , China, 8-12 Aug. 2005.
[15] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architecture
for a ubiquitous RFID sensing network,” in 2nd International Conference on Intelligent Sensors, Sensor
Networks and Information Processing (ISSNIP), Melbourne, Australia, 5-8 Dec. 2005.
[16] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architec-
ture for a ubiquitous item identification network,” in Seventh International Conference on Ubiquitous
computing, Tokyo, Japan, 11-14 Sept. 2005.
Page xxiv
Publications
Non-refereed
[1] K. S. Leong, and M. L. Ng, “A simple EPC enterprise model,” in Auto-ID Labs Workshop, Zurich, 23-24
Sept. 2004.
[2] K. S. Leong, M. L. Ng, and D. W. Engels, “EPC network architecture,” in Auto-ID Labs Workshop,
Zurich, 23-24 Sept. 2004.
[3] M. L. Ng, K. S. Leong, and D. W. Engels, “Prospects for ubiquitous item identification,” in Auto-ID
Labs Workshop, Zurich, 23-24 Sept. 2004.
Page xxv
Page xxvi
List of Figures
1.1 A simple RFID system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 The complete structure of the thesis . . . . . . . . . . . . . . . . . . . . . 7
2.1 RFID system: reader and tag relationship . . . . . . . . . . . . . . . . . . 14
2.2 The principle of backscattering . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Early stage EPC Network . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4 The Current EPC Network . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.1 Experimental path loss results . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 Plot of proposed piece-wise linear in-building path loss model against
distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.3 A simple illustration of Reader Interference with Tag Replies (RITR) . . 35
3.4 Polar plot of the antenna gain of a directional circularly polarised RFID
antenna with a gain of 6 dBi . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.5 Transmit mask for multiple-interrogator and dense-interrogator envi-
ronments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.6 Grid used in simulation software . . . . . . . . . . . . . . . . . . . . . . . 45
3.7 The orientation and the position of a transmitting antenna with respect
to the simulation grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.8 Simulation results verifying the functionality of the software developed
in MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.9 Common positionings of antennas with respect to each other . . . . . . . 51
3.10 Simulation results on one antenna simulation . . . . . . . . . . . . . . . . 53
3.11 Simulation results of a two antenna simulation . . . . . . . . . . . . . . . 56
Page xxvii
List of Figures
3.12 The configuration of a horizontal antenna on ground or facing the ground
with respect to the normal configuration of a transmitting antenna . . . 57
3.13 Simulation results showing the effect of horizontal antenna configuration 58
3.14 Simulation results for some antennas operating in different channels and
pointing in different directions . . . . . . . . . . . . . . . . . . . . . . . . 60
3.15 Determination of safe distance in accordance to LBT when the second
antenna is an isotropic radiator . . . . . . . . . . . . . . . . . . . . . . . . 61
3.16 Repositioning of a horizontal antenna . . . . . . . . . . . . . . . . . . . . 64
3.17 Reader synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.1 Equipment used in path loss prediction experiment . . . . . . . . . . . . 73
4.2 Experiment grid for signal strength measurement . . . . . . . . . . . . . 74
4.3 Measuring configuration from side view . . . . . . . . . . . . . . . . . . . 78
4.4 Plot of path loss using different models together with measured results. 81
4.5 Comparison of the received signal strength with and without the pres-
ence of a nearby metallic box . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.6 Polar plot of the antenna gain of a directional circularly polarised RFID
antenna with a gain of 6 dBi . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.7 Boundary zone in simulation results . . . . . . . . . . . . . . . . . . . . . 86
4.8 Second carrier sensing with “Listen Before Talk” provision . . . . . . . . 88
4.9 Common antenna positioning . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.10 Arrangement of 64 reader antennas for SAR simulation . . . . . . . . . . 98
4.11 The experiment site of the 64 reader antenna experiment . . . . . . . . . 99
4.12 Plot of simulation results on SAR investigation . . . . . . . . . . . . . . . 102
5.1 Synchronisation of all readers . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.2 Centralised LBT system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Page xxviii
List of Figures
5.3 Localised LBT system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.4 Antenna positioning in dock door . . . . . . . . . . . . . . . . . . . . . . 113
5.5 Alternating of “Listening” and “Talking” mode . . . . . . . . . . . . . . . 115
5.6 Channelling of the allocated frequency spectrum . . . . . . . . . . . . . . 115
5.7 Reduction of output power to fine-tune reader synchronisation . . . . . 116
5.8 Using sensors in an RFID system . . . . . . . . . . . . . . . . . . . . . . . 117
5.9 Use of RF absorbing materials . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.10 Channel switching within antennas . . . . . . . . . . . . . . . . . . . . . . 119
5.11 The complete frequency band allocated for RFID operation in Europe . . 120
5.12 Separation of transmitting and receiving channels in the frequency band
allocated for RFID operation in Europe . . . . . . . . . . . . . . . . . . . 120
6.1 The difference between source and vortex . . . . . . . . . . . . . . . . . . 126
6.2 Boundary conditions affecting electric and time varying magnetic field . 127
6.3 Spherical coordinate system . . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.4 Series resonant circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.5 Parallel resonant circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.6 Practical parallel resonant circuit . . . . . . . . . . . . . . . . . . . . . . . 139
6.7 Reflection coefficient for best utilisation of λRC . . . . . . . . . . . . . . . . 142
6.8 The proposed chip bonding on to a thin metal strip . . . . . . . . . . . . 144
6.9 Reuse of HF RFID tag chip . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.10 A ISD72128 or C220 HF RFID tag . . . . . . . . . . . . . . . . . . . . . . . 145
6.11 72128 TTF chip on smart card module . . . . . . . . . . . . . . . . . . . . 146
6.12 A Texas Instruments UHF strap . . . . . . . . . . . . . . . . . . . . . . . . 146
6.13 An Alien C1G2 UHF strap . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.14 FEIG HF RFID Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.15 Gemplus HF RFID reader . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Page xxix
List of Figures
6.16 FEIG UHF RFID reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.17 Alien UHF RFID reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.18 Typical HF antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.19 Effect of using Z-axis conductive tape on an HF RFID tag . . . . . . . . . 151
6.20 PML layer in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.21 Linking MATLAB and HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . 156
6.22 The auto generated loop using VBS . . . . . . . . . . . . . . . . . . . . . . 157
6.23 Plots |S11| curve in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.24 Plots in HFSS showing the normalised |S11| curve . . . . . . . . . . . . . 159
6.25 Computation of the mutual inductance between two planar wire segments165
6.26 A Zork wine closure model . . . . . . . . . . . . . . . . . . . . . . . . . . 167
6.27 RFID tag for wine closure . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
6.28 Casing for livestock RFID tag . . . . . . . . . . . . . . . . . . . . . . . . . 170
6.29 Tag encapsulation casing from Leader Products . . . . . . . . . . . . . . . 170
6.30 First version of a pig tag with bent tracks to connect all the circular tracks
together. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.31 Second version of a pig tag using a spiral track . . . . . . . . . . . . . . . 171
6.32 First prototype of a pig tag . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
6.33 Problem of uneven track thickness when prototyping antenna with thin
tracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
6.34 First version prototype of a pig tag . . . . . . . . . . . . . . . . . . . . . . 173
6.35 Pig tag with TTF compatibility . . . . . . . . . . . . . . . . . . . . . . . . 174
6.36 Testing a pig tag with TTF compatibility . . . . . . . . . . . . . . . . . . . 174
6.37 Pig tag with external capacitance added . . . . . . . . . . . . . . . . . . . 176
6.38 Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of
1.6 mm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
6.39 Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of
0.8 mm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Page xxx
List of Figures
6.40 Final version prototype of a pig tag . . . . . . . . . . . . . . . . . . . . . . 177
6.41 HF RFID Tag to be embedded in a livestock ear tag . . . . . . . . . . . . 178
6.42 RFID tags before and after encapsulation process . . . . . . . . . . . . . . 179
6.43 An encapsulated tag immersed in water to test for readability and water
proof capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
6.44 Overall RFID system setup in piggery . . . . . . . . . . . . . . . . . . . . 180
6.45 HF and UHF RFID readers in piggery . . . . . . . . . . . . . . . . . . . . 180
6.46 Orientation of HF and UHF tags on pigs ears . . . . . . . . . . . . . . . . 180
6.47 Picture on pigs feeding in an RFID pig feeder . . . . . . . . . . . . . . . . 181
6.48 Sheep ear tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
6.49 Two UHF pig tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
6.50 Structure of the RFID tag with a rectangular loop antenna . . . . . . . . . 187
6.51 RFID tag for wine closure . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
7.1 A generic HF coil antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
7.2 A generic UHF dipole antenna . . . . . . . . . . . . . . . . . . . . . . . . 197
7.3 First version of a dual-frequency antenna . . . . . . . . . . . . . . . . . . 199
7.4 Impedance of the first version of dual-frequency antenna at HF . . . . . 199
7.5 Impedance of the first version of dual-frequency antenna at UHF . . . . 200
7.6 Second version of dual-frequency antenna . . . . . . . . . . . . . . . . . . 201
7.7 Impedance of the second version of dual-frequency antenna at UHF . . 202
7.8 Graph showing the calculated impedance and simulated impedance of
the second version of dual-frequency antenna . . . . . . . . . . . . . . . . 205
7.9 Location of feed point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
7.10 Smith Chart showing the simulated impedance of the dual-frequency
antenna with different transmission line length . . . . . . . . . . . . . . . 207
7.11 A UHF dipole with a matching network . . . . . . . . . . . . . . . . . . . 208
Page xxxi
List of Figures
7.12 A UHF dipole with tuned matching network . . . . . . . . . . . . . . . . 208
7.13 Impedance of the UHF dipole with tuned matching network . . . . . . . 209
7.14 A UHF dipole with bent sides . . . . . . . . . . . . . . . . . . . . . . . . . 210
7.15 Impedance of the UHF dipole with bent sides . . . . . . . . . . . . . . . . 210
7.16 A UHF dipole with bent sides and gap . . . . . . . . . . . . . . . . . . . . 211
7.17 A UHF dipole with bent sides and repositioned gap . . . . . . . . . . . . 211
7.18 Final version of the UHF dipole, with matching network, bent sides and
gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
7.19 HF Impedance of the final version of the UHF dipole shown in Fig 7.18 . 213
7.20 UHF Impedance of the final version of the UHF dipole shown in Fig 7.18 213
7.21 Self resonance frequency of HF planar coil antenna shown in Fig. 7.6 . . 214
7.22 Merging of HF multi-turn loop antenna with a UHF dipole . . . . . . . . 215
7.23 Final design for dual-frequency antenna . . . . . . . . . . . . . . . . . . . 216
7.24 Simulated impedance for the dual-frequency antenna at HF . . . . . . . 217
7.25 A transmission measurement using a wide-band loop antenna . . . . . . 218
7.26 Fabricated final version of the designed dual-frequency antenna . . . . . 218
7.27 Transmission measurement using network analyser to locate the reso-
nance point at HF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
7.28 The simulated and measured results for the frequency response of the
fabricated dual-frequency antenna within the RFID UHF band from 860
- 960 MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
7.29 A different version of a dual-frequency antenna where the dipole is lo-
cated in the inside of the HF coil . . . . . . . . . . . . . . . . . . . . . . . 222
7.30 The HF impedance of the dual-frequency antenna in Fig 7.29 . . . . . . . 222
7.31 The UHF impedance of the dual-frequency antenna in Fig 7.29 . . . . . . 223
7.32 A miniaturised HF coil antenna . . . . . . . . . . . . . . . . . . . . . . . . 223
7.33 A miniaturised UHF dipole . . . . . . . . . . . . . . . . . . . . . . . . . . 224
7.34 A dual-frequency antenna with miniaturised HF coil and UHF dipole . . 225
Page xxxii
List of Figures
7.35 Impedance transformation in Smith Chart where all the traces cover 860
to 960 MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.36 Dual-frequency antenna with rotated HF coil . . . . . . . . . . . . . . . . 228
7.37 Simulated results at HF for antenna in Fig. 7.36 . . . . . . . . . . . . . . . 229
7.38 Simulated results at UHF for antenna in Fig. 7.36 . . . . . . . . . . . . . . 229
7.39 Figure to illustrate the position of RFID chip attached to the antenna . . 230
7.40 A miniaturised dual-frequency antenna . . . . . . . . . . . . . . . . . . . 232
7.41 Final antenna design with design parameters . . . . . . . . . . . . . . . . 232
8.1 Feed point repositioning of the dual-frequency antenna . . . . . . . . . . 237
8.2 Impedance of dual-frequency antenna with feed point at position A at HF238
8.3 Impedance of dual-frequency antenna with feed point at location A at
UHF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
8.4 A UHF patch antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
8.5 The HF antenna to be fine-tuned to operate at UHF . . . . . . . . . . . . 242
8.6 Simulated impedance of the HF antenna shown in Fig. 8.5 at HF . . . . . 242
8.7 Simulated impedance of the HF antenna shown in Fig. 8.5 at UHF . . . . 243
8.8 The structure of the HF antenna shown in Fig. 8.5 with illustration of
surface current density on the top side . . . . . . . . . . . . . . . . . . . . 244
8.9 Measurement result of the HF antenna shown in Fig. 8.5 at UHF . . . . . 246
8.10 Stacking of HF and UHF RFID tags . . . . . . . . . . . . . . . . . . . . . . 248
8.11 Frequency response of the HF tag at UHF without the presence of any
UHF tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
8.12 Frequency response of the UHF tag at UHF when placed near to HF tag 250
8.13 Frequency response of the UHF tag at UHF when the tag is able to be
read by an interrogator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Page xxxiii
List of Figures
9.1 Explanation of common mode current . . . . . . . . . . . . . . . . . . . . 258
9.2 Cables used for measuring experiment . . . . . . . . . . . . . . . . . . . . 260
9.3 A bow tie antenna on FR4 . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
9.4 Half bow tie antenna on ground plane . . . . . . . . . . . . . . . . . . . . 261
9.5 The effect of SMA connector on measurement results . . . . . . . . . . . 262
9.6 Simulated and measured resistance of AUT . . . . . . . . . . . . . . . . . 263
9.7 Simulated and measured reactance of AUT . . . . . . . . . . . . . . . . . 263
9.8 The designed Balun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
9.9 The power transfer efficiency between AUT and chip when reactances
are equal and opposite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
9.10 The power transfer efficiency between AUT and chip when resistances
are equal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
9.11 The comparison between the estimated and calculated values on the
power loss curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
B.1 Illustration to explain the calculation of flux density of a round wire of
radius a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
F.1 Antennas on tripods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
F.2 Measurement of signal strength . . . . . . . . . . . . . . . . . . . . . . . . 306
F.3 Path taken when measuring signal strength from the west side of Engi-
neering North building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
F.4 Path taken when measuring signal strength from the east side of Engi-
neering North building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
F.5 Path taken when measuring signal strength from Engineering and Maths
(EM) building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
F.6 Comparison between calculated and measured received signal strength
(in building) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Page xxxiv
List of Figures
F.7 Antennas facing directly perpendicular to the wall . . . . . . . . . . . . . 311
F.8 Antennas facing with an angle to the wall . . . . . . . . . . . . . . . . . . 311
F.9 Antennas on tripods facing a conductive fence . . . . . . . . . . . . . . . 311
F.10 Measurement of signal strength in outdoors . . . . . . . . . . . . . . . . . 313
F.11 Antennas on the ground and facing upwards . . . . . . . . . . . . . . . . 313
G.1 A pig feeder used in RFID deployment experiment . . . . . . . . . . . . 316
G.2 Modelling of a pig feeder (Stage 1) . . . . . . . . . . . . . . . . . . . . . . 317
G.3 Modelling of a pig feeder (Stage 2) . . . . . . . . . . . . . . . . . . . . . . 318
G.4 Overview arrangement of RFID deployment in piggery . . . . . . . . . . 323
G.5 The position of an antenna in an antenna box besides a pig feeder . . . . 324
G.6 Encapsulation of a HF reader . . . . . . . . . . . . . . . . . . . . . . . . . 325
G.7 Encapsulation of a UHF reader . . . . . . . . . . . . . . . . . . . . . . . . 326
G.8 Network setup in a piggery . . . . . . . . . . . . . . . . . . . . . . . . . . 326
Page xxxv
Page xxxvi
List of Tables
3.1 The effect of tag distance on multi-reader interference . . . . . . . . . . . 36
3.2 UHF RFID reader radiated power with corresponding threshold values
for LBT, and minimum allowable distance between antennas . . . . . . . 42
3.3 Safe distances, in m, for different antenna positioning to avoid the Reader
Interference with Tag Replies (RITR) problem . . . . . . . . . . . . . . . . 52
3.4 Safe distance for different antenna positionings to avoid Regulatory Reader
Shutdown (RRS) problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.5 Safe distance for different antenna configuration when the second an-
tenna is an isotropic radiator in accordance to LBT . . . . . . . . . . . . . 62
4.1 UHF RFID reader effective radiated power with corresponding thresh-
old values for LBT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.2 Average prediction of path loss based on experimental results . . . . . . 75
4.3 Difference in wavelength between direct distance, r, and reflected path, D 79
4.4 Comparison of measured and predicted signal strength . . . . . . . . . . 84
4.5 Threshold values for second carrier sensing with respect to tag read
range for different antenna orientations . . . . . . . . . . . . . . . . . . . 92
4.6 Threshold values for second carrier sensing with respect to environment
factor, n, for different antenna orientations . . . . . . . . . . . . . . . . . 93
4.7 Simulation results on SAR investigation using maximum power for ev-
ery channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.8 Calculation results of the worst-case scenario . . . . . . . . . . . . . . . . 101
5.1 UHF RFID reader radiated power with corresponding threshold values
for LBT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.2 Safe distance for different antenna configuration when the second an-
tenna is an isotropic radiator in accordance to LBT . . . . . . . . . . . . . 109
Page xxxvii
List of Tables
6.1 SI units for common terms used in antenna theory . . . . . . . . . . . . . 125
6.2 Inductance values for spiral planar coil with different numbers of turns,
with L = 20; L1=40; w=3; g=1 . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.3 Inductance values for various spiral planar coils . . . . . . . . . . . . . . 166
6.4 Performance of HF RFID tag for Zork . . . . . . . . . . . . . . . . . . . . 168
6.5 Read range for UHF tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
6.6 Read range for UHF tag for beer keg . . . . . . . . . . . . . . . . . . . . . 187
6.7 Performance of UHF RFID tag for the Zork wine closure . . . . . . . . . 188
7.1 HF versus UHF in RFID operation . . . . . . . . . . . . . . . . . . . . . . 193
7.2 Calculation of combined impedance based on individual impedances . . 203
7.3 Comparison between calculated impedances and simulated impedances 204
8.1 Measurement results for the UHF read range of the dual-frequency an-
tenna using the stacking method . . . . . . . . . . . . . . . . . . . . . . . 249
8.2 Comparison between different methodologies in creating a dual-frequency
RFID tag antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
9.1 Average error and maximum error . . . . . . . . . . . . . . . . . . . . . . 266
F.1 Measured and calculated signal strengths for various locations . . . . . . 309
F.2 Measured signal strength for antennas facing a wall or a conductive fence 312
F.3 Calculated and measured signal strength (outdoor) for different antenna
orientations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
G.1 Performance of HF RFID tag before and after encapsulation . . . . . . . 320
G.2 Performance of UHF RFID tag before and after encapsulation . . . . . . 321
G.3 HF and UHF RFID tags attached to pigs . . . . . . . . . . . . . . . . . . . 327
G.4 Results from remote monitoring . . . . . . . . . . . . . . . . . . . . . . . . 330
Page xxxviii
Chapter 1
Thesis Introduction
THIS chapter explains the title of this thesis, which is “Antenna
Positioning Analysis and Dual-Frequency Antenna Design of
High Frequency Ratio for Advanced Electronic Code Responding
Labels”. The background of the research topic, Radio Frequency Identifica-
tion (RFID), is presented. The motivation and contribution to this research
is discussed, followed by the elaboration of the structure of this thesis.
Page 1
1.1 Thesis Title
1.1 Thesis Title
The title of this thesis is “Antenna Positioning Analysis and Dual-Frequency Antenna
Design of High Frequency Ratio for Advanced Electronic Code Responding Labels”.
There are two related areas of research presented in this thesis. The first area is “An-
tenna Positioning Analysis”, while the second area is “Dual-Frequency Antenna De-
sign of High Frequency Ratio for Advanced Electronic Code Responding Labels”.
In “Antenna Positioning Analysis”, the term “Antenna” refers to the Radio Frequency
Identification (RFID) reader’s antenna. The positioning analysis done is on the place-
ment of RFID readers’ antennas to minimise readers’ interference, to offer more cover-
age while conforming to local regulations. The analysis includes but is not limited to
the investigation of in-door path loss models, the computation of signal strength, and
the examination of Specific Absorption Rate (SAR).
“Dual-Frequency Antenna Design of High Frequency Ratio for Advanced Electronic
Code Responding Labels” involves the design of an RFID tag antenna and is not re-
lated to the RFID reader antenna. Frequency ratio is computed through the division
of the higher frequency by the lower frequency and hence is always greater than 1. A
dual-frequency antenna with low frequency ratio (<5) is common, but not for a dual-
frequency antenna with high frequency ratio. The two frequencies of interest in this
thesis are the HF band (13.56 MHz) and the UHF band (860 - 960 MHz) and hence the
intended dual-frequency antenna has a frequency ratio in excess of 70. The work pre-
sented in this thesis includes the design of individual HF tag antennas and UHF tag
antennas, as will be shown in detail in Section 1.5.
1.2 Background of Research
Radio Frequency Identification (RFID) forms the background of this research. An RFID
system, is in fact a very complex system, which involves various fields of study, such as
antenna design, signal processing, integrated circuit design, RF hardware design and
information networking. The focus will be on an RFID system for the supply chain
Page 2
Chapter 1 Thesis Introduction
to track individual items, and thus focuses on either HF or UHF RFID systems using
passive RFID tags.
A simple RFID system is as shown in Fig. 1.1. An RFID tag, which contains a unique
Electronic Product Code (EPC) [4] is attached to an item. Since the tag is passive, an
RFID reader antenna is required to power up the tag. Once the RFID reader detects the
tag, the EPC of the tag is sent to a host computer for processing and business decision
making. For example, if product A is checked out at a sales point, the host computer
will receive the EPC of Product A and when the stock of Product A is running low, the
host computer can automatically generate a purchase order of Product A. A detailed
review of RFID systems is presented in Chapter 2.
RFID Network(EPC Network)
RFIDReader
Antenna RFIDTag
HostComputer
1 2
3
Figure 1.1. A simple RFID system. (1) Host computer to link RFID readers to an RFID Network
(generally an Electronic Product Code (EPC) Network). (2) An RFID reader with one or
more reader antennas connected. (3) Passive RFID tags powered up by reader antennas
so that these tags can be read by the readers.
The research areas involved, according to the societies of the Institute of Electrical and
Electronics Engineers (IEEE), are [5]:
• Antennas and Propagation
• Electromagnetic Compatibility
• Microwave Theory and Techniques
1.3 Motivation
RFID is not a new technology. As will be discussed in Chapter 2, the history of RFID
can be traced back to the early 20th century [6]. However, it is only in the late 1990’s
Page 3
1.3 Motivation
that RFID gained widespread attention when its potential in automating the supply
chain was investigated. It is believed that RFID can revolutionise the supply chain
using real time monitoring, which can offer instant restocking capabilities, elimination
of the overstocking problem, reduction in counterfeiting and many others.
To fully embrace the potential of an RFID system in a supply chain, it is necessary
to attach an RFID tag to every box, carton or pallet in the supply chain. Also, every
RFID tag contains a chip which holds a unique identification code of the object the
tag is attached to. Hence only passive RFID tags are feasible to be deployed in this
manner, or else the total cost of any product with an RFID tag attached will increase to
an unacceptable level.
As mentioned before, the focus of this research is on the deployment of RFID systems
in the supply chain. A supply chain poses a challenging environment for the deploy-
ment of passive RFID tags. Firstly, there are physically thousands to millions of RFID
tags in a relatively small area, such as a warehouse. In other words, a dense RFID
tag environment is common along a supply chain. Intensive research needs to be car-
ried out to ensure good RFID coverage, reliable RFID systems, minimising interference
between tags and readers, while optimising the frequency spectrum in accordance to
local regulations.
Secondly, in a conventional passive RFID system, the reader antenna provides power
to the passive RFID tags before any communication link between a reader and a tag
can be established. The quality of this communication link depends heavily on the
environment. In the worst case scenario, a hostile environment can render an other-
wise functioning RFID system useless. One very interesting point is that ionised liquid
products are very common along the supply chain. Ionised liquid, such as water, can
absorb microwave power easily and hence a UHF RFID system will not perform well
in an ionised liquid filled environment. This prompts an investigation to design a dual-
frequency RFID tag antenna, which can operate at UHF in normal operation and also
when desired at HF when ionised liquid is present.
Page 4
Chapter 1 Thesis Introduction
1.4 Contribution
There are two main areas of contribution offered by this thesis with respect to the two
motivations as discussed in Section 1.3.
The first area of contribution is on the dense reader deployment. This thesis presents a
solid and practical way to perform reader antenna positioning analysis. The research
started off with the study of a simple path loss model. This simple model was then
modified to take into consideration environmental factors to cater for different deploy-
ment situations. With a comprehensive model, simulation code was written in MAT-
LAB to predict the effect of RFID readers on each other. This offers an easy way to
examine the interference between readers. Though approximate, as the actual envi-
ronment is dynamic in nature, this allows examination of potential interference before
actual implementation is carried out.
Also, this thesis investigates some novel techniques in synchronising the operation
of RFID readers in a certain vicinity to optimise the performance of an RFID system
when operating under strict local regulations. The local regulation of interest is the
European “Listen Before Talk” provision. The author believes that “Listen Before Talk”
is not beneficial for the widespread adoption of RFID system. Intensive studies carried
out proved that the threshold values as specified in the “Listen Before Talk” provision
are too low to allow wide spread adoption of RFID to be feasible. It is hoped that
the results presented here will strengthen the case for the relaxation of the regulation
imposed upon RFID systems in Europe.
Though not entirely related to those issues discussed above, this thesis extends its
reader positioning study into the implementation of a new technique, named “Sec-
ond Carrier Sensing”. This technique is to prevent the RFID tag from being confused
by multiple interrogation signals from different reader antennas which exist in one
area. Also, included is the examination of Specific Absorption Rate (SAR) in a dense
reader environment. The analysis was carried out using both the path loss model de-
veloped and the worst case scenario, to ensure the deployment of RFID readers would
not create any potential radiation overexposure to anyone.
Page 5
1.4 Contribution
The second area of contribution is concerned with the art of RFID tag antenna de-
sign. This thesis presents the crafting of a novel RFID tag antenna from very basic HF
and UHF tag antennas. The design specifications of the tag antennas in this thesis are
mostly industry driven, where various actual implementations are possible (some were
already implemented in the course of this research). Design decisions are clearly pre-
sented to offer a reference for those interested in tag design for other applications. The
work on individual HF and UHF RFID tag antennas is essential in providing a strong
foundation for the research on a novel dual-frequency antenna, as a dual-frequency
antenna is in fact an antenna with the characteristics of both HF and UHF antennas.
With this foundation, the research on tag antenna design was then extended into the
design of a dual-frequency RFID tag antenna. This dual-frequency antenna supports
both HF band and UHF band operation and has a frequency ratio of more than 70.
This is the first dual-frequency antenna that supports such a high frequency ratio while
maintaining a single feed structure. With this dual-frequency antenna, an RFID tag can
support two very different frequency bands with just a single chip. This is especially
important as HF will offer a better read range in a environment with lots of liquid
products. Indirectly, this will contribute to the effort in spreading and encouraging the
usage of RFID in the supply chain.
Alternative methods in creating a novel dual-frequency antenna of similar character-
istics are also explored, with the advantages and disadvantages of each method being
discussed and compared. This provide a clear picture on the design options of a dual-
frequency antenna for an RFID tag.
Also documented in this thesis are the techniques developed over the the entire re-
search period for measuring the performance of newly designed RFID tag antennas.
This will provide a compact guide for anyone who is interested in the area of RFID tag
design.
A summary of the contribution can be found in the conclusion of this thesis in Sec-
tion 10.4.
Page 6
Chapter 1 Thesis Introduction
1.5 Thesis Structure
This thesis is structured into 9 main chapters, excluding conclusion, appendix and
bibliography, and is as shown in Fig. 1.2.
Chapter 2RFID System
Chapter 3RFID Positioning
Analysis
Chapter 5Reader
Synchronisation
Chapter 4Operational
Considerations
Chapter 6RFID Antenna
Design
Chapter 7Dual-frequencyRFID Antenna
Chapter 8Alternative Dual-frequency
RFID Antenna
Chapter 9Antenna
Measurement
Figure 1.2. The complete structure of the thesis.
These 9 chapters can be categorised into 3 major areas:
1. Introduction
• Chapter 1 of this thesis (which is this chapter) explains the title of this thesis
and outlines the background of this research. Also highlighted is the moti-
vation of this research and its contribution to the engineering community.
• Chapter 2 presents a survey of RFID systems and includes a brief history
of RFID, a presentation of a simple RFID system and of variants of RFID
systems. Also discussed are the deployment of RFID in the supply chain,
RFID standards and the future of RFID.
2. Antenna Positioning Analysis
• Chapter 3 is on the RFID reader antenna positioning analysis. Simple path
loss models are discussed and applied in the prediction of the strength of
Page 7
1.5 Thesis Structure
RFID reader interrogating signals. Furthermore, a simulation coded in MAT-
LAB was developed.
• Chapter 4 investigates the operational considerations for RFID deployment
and simulation. It extends the research presented in Chapter 3 and offers
guidelines to avoid pitfalls in setting up simulation software for RFID reader
deployment. Furthermore, the idea of “Second Carrier Sensing” is presented
with its limits discussed in relation to the positions of reader antennas. An
in-depth investigation of Specific Absorption Rate (SAR) is also included,
which is essential before any deployment of dense RFID system.
• Chapter 5 presents the idea of RFID reader synchronisation, where RFID
readers are synchronised to share the limited bandwidth allowed under cer-
tain local regulations. A variation of the synchronisation method is intro-
duced to cater for different needs in real life implementation.
3. Dual-Frequency RFID Tag Antenna Design and Analysis
• Chapter 6 is a large chapter, and mainly deals with the design of RFID tag
antennas. In the beginning of the chapter, the design constraints of small
RFID tag antennas are discussed. The latter parts can be separated into two
major parts, the design of HF RFID tag antennas and the design of UHF
RFID tag antennas.
In the design of HF RFID tag antennas, the theory of a simple HF loop an-
tenna (magnetic dipole) is reviewed. This is followed by the practical design,
simulation, fabrication and testing of HF RFID tags for various applications
which include wine cork tagging and pig ear tagging.
In keeping with above, in the design of UHF RFID tag antennas, the theory
of the simple electric dipole and in particular of its planar form is reviewed,
followed by the practical design, simulation, fabrication and testing of UHF
RFID tags for various applications which include sheep ear tagging and beer
keg tagging.
• Chapter 7 focuses on the design and analysis of a dual-frequency RFID tag
antenna. Though different methodologies are presented, the main focus
Page 8
Chapter 1 Thesis Introduction
of this chapter is to design a dual-frequency antenna using the method of
merging an HF antenna and a UHF antenna. Theories for antenna design
were reviewed before actual design, fabrication, and actual tag testing was
carried out. Also included in this chapter, is the development of a minia-
turised version of the designed dual-frequency RFID tag antenna.
• Chapter 8 complements Chapter 7 and explores alternative methods in cre-
ating a dual-frequency antenna, which includes having different feed point
locations or different antenna types when merging an HF together with a
UHF antenna. Furthermore, research is extended into having an HF antenna
acting as a UHF antenna at the UHF band and having a UHF antenna acting
as an HF antenna at the HF band.
• Chapter 9 discusses the measurement techniques used in carrying out all
the measurements to obtain the characteristics of the RFID tag antenna. The
focus was on the technique in obtaining the correct impedance of a UHF tag
antenna. Also included is the analysis on the power transfer between a tag
antenna and the attached chip, when there is mismatch caused by inaccu-
racy in impedance measurement.
1.6 Chapter Structure
A chapter begins with a brief introduction, followed by background study of materials
related to the research and discussion of that chapter. Chapter 2 will be referred to if
any basic RFID background is involved. Most of the important results in any chapter
have been published as conference papers. Any publication used in a chapter will be
mentioned at the end of the introduction section of that chapter or the beginning of a
section.
Page 9
Page 10
Chapter 2
RFID Systems
THIS chapter presents a study of the fundamental principles of
operation of a passive RFID system. This study and the asso-
ciated literature review provide a strong foundation for the re-
search carried out by the author and will be referred to throughout this
thesis.
Page 11
2.1 Introduction
2.1 Introduction
Radio Frequency Identification (RFID) is a technology which involves several distinct
areas of expertise. As explained in Chapter 1, the focus of research presented in this
thesis is on the RFID antenna positioning analysis in a dense reader deployment zone,
and the antenna design for a dual-frequency antenna. Nonetheless, for completeness, a
detailed background on RFID is presented in this chapter, together with the operational
principles of RFID, that will be referred throughout this thesis.
Firstly, the history of RFID will be presented in the next section, followed by descrip-
tion of a simple RFID system in Section 2.3. Variants of RFID are discussed in Sec-
tion 2.4 though as mentioned in several occasions before, the focus of this thesis in on
passive HF and UHF RFID systems.
The attention is then focussed on the impact of RFID deployment in the supply chain
(Section 2.5) and the current network system responsible in handling the massive amount
of data created by RFID tag reading (Section 2.6). RFID standards and protocols, which
govern the operation of RFID system, are presented in Section 2.7.
Papers on the networking aspect of RFID have been co-authored [7, 8], and the key
points of those papers are duplicated in this chapter.
2.2 History of RFID
RFID, in fact, is not a new technology and has a long history. It is believed that the
birth of RFID can be traced back to the early 20th century, when Radio Detection And
Range (Radar) was invented, where the reflected wave is used to detect an object.
However, it was then discovered that detecting an object is not sufficient in some situ-
ation, especially during the war time, when an identification is also required. Hence,
the first RFID deployment was carried out as a long-range transponder systems of
“Identification, Friend, or Foe” (IFF) for aircraft, so that a friendly aircraft would not
be misidentified as the enemy [9, 6].
Page 12
Chapter 2 RFID Systems
The first patent on RFID is by Harris in 1960, titled “Radio transmission systems with
modulatable passive responder”. In this patent, Harris presented several ways of com-
munication between a reader and a passive responder (tag). [10].
Early journal publications on RFID include:
• “Small resonant scatters in field measurement” by Harrington [11] in 1962, where
a general formulation for a backscattered field from loaded objects is given.
• “Electromagnetic scattering by antennas” by Harrington [12] in 1963, where he
relates the theory of loaded scatterers to antenna theory.
• “Electronic surveillance system” by Cole [13] in 1972, where surface acoustic
waves (SAW) are utilised for the generation within a passive tag of a coded reply,
with electromagnetic communication between the tag and a reader antenna.
• “Identification using modulated RF backscatter” by Koelle, Depp and Freyman
[14] in 1975, where a successful demonstration of identification using RFID backscat-
tering is shown.
However, because of cost and size issues, RFID had limited usage and coverage in the
70’s and 80’s. It was deployed in certain toll collection and object tracking applications.
With the advances in semiconductor technology and miniaturisation in 90’s, the cost
and size of the passive transponders were reduced to a very acceptable level and at the
end of 20th century, they were beginning to be used in the supply chain.
2.3 A Simple RFID System
RFID is a technique used to identify objects by means of electromagnetic waves. An
object can be tagged with an electronic code responding label. An electronic tag con-
sists of an antenna and an integrated circuit. Upon receiving any valid interrogating
signal from any interrogating source, such as a reader, the tag will respond according
to its designed protocol. The relationship between a tag and a reader is illustrated as
Figure 2.1.
Page 13
2.4 Variants of RFID
Reader
Tag
Tag
Tag
TagAntenna
Forward LinkReturn Link
Figure 2.1. RFID system: reader and tag relationship. One or more antennas can be connected
to a reader, and are used to communicate with tags within the field of interrogation.
Signals sent from reader through an antenna to a tag are called forward link signals
while the signals from a tag back to reader are called return link signals.
A detailed description of an RFID system can also be found in RFID Handbook by
Finkenzeller [15].
2.4 Variants of RFID
2.4.1 Active and Passive RFID Tags
There are two types of tags, active and passive. Both active and passive RFID tags are
defined in [16].
An active tag is defined as an “RFID device having the ability of producing a radio
signal”. Normally an active tag has its own battery source; it has a greater read range
compared to a passive tag but is limited by the life time of its battery.
A passive tag is defined as an “RFID device which reflects and modulates a carrier
signal received from an interrogator”. A passive tag is normally energised by an in-
terrogating signal from a reader antenna and does not have any other internal power
source. It typically has shorter read range as compared with an active tag.
There are several good RFID passive tag chip designs available in the literature. For
example, a low power high performance UHF passive tag chip was presented by
Karthaus [17], which requires only 16.7 µW to power up the tag chip and has a read
range of around 10 m. There is also a low cost UHF passive tag design offered by
Page 14
Chapter 2 RFID Systems
Glidden [18]. Tag chips are mentioned here because knowledge of the tag chip input
impedance is essential in antenna design.
2.4.2 Backscattering
When an electromagnetic wave is radiated in a direction, part of it will be reflected
back to its source if there exists a reflective (metallic) object in its path of propagation.
This is the concept of radar system, where a foreign object in the airspace (such as an
airplane) can detected.
Backscattering is based on the similar concept. The only difference is that in backscat-
tering, the reflected wave is used to transmit data. By changing the impedance of the
reflective object, the reflected wave will be changed accordingly.
In the case of RFID operation, a passive RFID tag consists of two important parts,
the tag antenna and the tag chip. When this tag enters an interrogation zone of an
RFID reader antenna, it will be powered up by the reader antenna. When instructed to
backscatter its unique EPC number, the tag chip will change its input impedance in an
order representing its EPC in binary form (high impedance for “1” and low impedance
for “0”, or vice versa) while the reader is transmitting a continuous wave. The reflected
wave is monitored by the reader and then interpreted into the EPC number of the tag.
Zinput = A
Zinput = A’
Tag
Tag
F
F
R
R’
Figure 2.2. The principle of backscattering. The reflected wave changes depending on the input
impedance of the tag, and can be used to represent binary “1” and “0” to transmit
data. In the figure above, F is the forward wave, while R and R’ are distinguishable
reflected waves depending on the input impedance of the tag.
Page 15
2.4 Variants of RFID
2.4.3 Common Frequency Bands for RFID Operations
There are 4 popular bands in RFID, which are the LF band (less than 135 kHz), the
HF band (13.553 - 13.567 MHz and commonly known as the 13.56 MHz band), the
UHF band (860 - 960 MHz) and the so called microwave band (2.400 - 2.4835 GHz and
commonly known as the 2.45 GHz band) [19]. Each band has positive and negative
characteristics.
The Auto-ID Center has published protocols for HF and UHF RFID operations. LF
has been left out, as under the vision of Auto-ID Center, each tag has to bear an EPC
number, which must be at least 64-bits in length [4], and for low frequency operation,
such as in the LF band, very slow reading in a heavily populated tag environment
would result. Also, LF RFID tags require large antenna components and hence are
difficult to implement and are susceptible to electrical noise, which HF can handle
[20]. Microwave tags can offer comparatively very fast reading, but the performance
of microwave tags will suffer, worse than any other bands described above, in the
presence of liquid or metal [19]. It is very difficult to conclude whether HF or UHF
is better for RFID application in supply chains. This is because in any supply chain,
there are lots of different scenarios, and in some HF is better and in the others, UHF
outperforms HF operation.
The strength of UHF lies in the fact that it offers 4 - 5 m average read range (up to
10 m in state of art RFID systems) and is cheaper to produce in large volume [19]. UHF
tags run at a higher frequency and hence have a higher read rate [21]. In other words,
in a UHF system, more data can be transferred within a fixed period of time, and
hence UHF is often deployed in a situation where more on board data is required [22].
However UHF RFID operation has difficulties in environments that are packed with
liquid or contain lots of metal, or even will not work properly in livestock industries
such as in the area of feedlots or slaughterhouses [23]. This is caused by the nature of
the radio frequency spectrum absorption by water as shown by [24], in which UHF is
highly absorbable.
On the other hand in item level tagging, HF has shorter read range (around 1 m) and
has a slower tag discrimination speed as compared to UHF [22], though it performance
Page 16
Chapter 2 RFID Systems
is still acceptable with a read rate of more than 200 tags per second. However, HF can
penetrate water better than UHF. According to the radio frequency spectrum absorp-
tion by water shown by [24], the absorption of HF by water is insignificant as com-
pared to UHF. Ref. [21] supports the claim of HF can penetrate water better because of
its longer wave length as compared to UHF, though HF also suffers in the presence of
metal just like UHF. Schurmann also reported that lower frequency HF systems are pri-
marily used in the case where better penetration of objects is needed as quoted by [15].
As a general consensus, both HF and UHF RFID systems are applicable in supply
chains, depending on the RFID application.
2.5 RFID and Supply Chain
As mentioned before, the deployment of RFID systems in supply chains in large num-
bers, with the potential to revolutionise supply chains, has sparked public interest
in RFID technology. A lot of case studies have been carried out by market analysts
[19, 22, 25, 26, 27, 28, 29, 30, 31]. Most of these articles discuss the impact of RFID on
the current supply chain, and how a company can maximise the benefits of an RFID
system.
The formation of the Auto-ID Center in 1999 aimed to develop an open standard ar-
chitecture for creating a seamless global network of physical objects [32]. The ulti-
mate vision was to create an “Internet of things”, and the concepts were well ex-
plained by Cole [25]. With the standardisation of HF protocols (ISO standardisation for
13.56 MHz: ISO/IEC 18000-3 [33] and ISO/IEC 15693 [34]) and UHF protocols (EPC-
global standardisation: UHF Class 1 Generation 2 [35] and EPC tag data standard [4]),
major parties in supply chains were pushing for RFID adoption. For example, one of
the biggest retail stores in the USA, Walmart, mandated the use of RFID in its supply
chains by its top 100 suppliers by Jan. 2005 [36]. Also, the USA Department of Defense
mandated the use of passive RFID in its supply chains by Jan. 2005 [37].
According to a Georgia Institute of Technology business research report, supply chain
problems cost companies between 9 and 20 percent of their value over a six-month
Page 17
2.5 RFID and Supply Chain
period [27]. This shows that the operation of the supply chains cannot afford to be
compromised, and any operational mistake would be costly.
Currently, inventory checking and tracing in a supply chain is done by reading bar-
codes. A barcode, explained in [15], is a binary code comprising a field of bars and
gaps arranged in a parallel configuration. It is read by optical laser scanning, i.e. by
the different reflection of a laser beam from the black bars and white gaps. In other
words, barcode scanning requires manual handling and orientation of each object and
also requires line-of-sight.
A thorough comparison between barcodes and RFID is offered by [15] and supported
by various other independent sources [19, 21, 38, 29]. They claim that, since RFID
does not require line-of-sight, scanning is faster, more effective and hence saves cost.
Furthermore, RFID identifies a unique product item, and enhances Warehouse Man-
agement System (WMS) in dealing with stock replenishment, over-stocking problems
and other supply chain related problems such as counterfeiting.
For example, in an international pharmaceutical supply chain, seven percent of drugs
may be counterfeit. Not only could RFID assure a safer supply of drugs to the con-
sumers, RFID could also save up to 8 billion dollars by 2006 [21]. This is something
that cannot be achieved by barcodes. The USA Food and Drug Administration (FDA)
trusts RFID as a tool to help authenticate drugs, and has taken the initiative in set-
ting up guidelines in using an RFID system in their pharmaceutical supply chain [39].
Also, business cases were presented to support RFID deployment and it is believed
that RFID will offer near perfect inventory information visibility in supply chains [26].
An RFID tag cost as low as 5 cents has been projected [40]. This is also the required cost
of a tag, so that every object in supply chains can be tagged with an RFID tag. Although
this is not the case currently, it is believed that the cost of an RFID tag can be reduced to
that level when the RFID systems are deployed in large numbers and across multiple
applications so that RFID tag production volume is increased to billions of tags.
One of the biggest challenges in RFID large scale deployment is the consumer privacy
issue. Though this is not included in this research, it is mentioned here to show the
author’s belief in RFID policies to avoid any individual privacy invasion.
Page 18
Chapter 2 RFID Systems
There are many reports on consumers’ fears of RFID deployment in supply chains [22].
Since RFID scanning is a type of automatic scanning [29], or in other words, a person
holding an item will not know when that item is being scanned, consumers may fear
that they are being tracked unlawfully.
The current RFID technology strives to address this issue. The measures taken include
public education on RFID based on the facts that RFID does not have long range scan-
ning capability, and the option that a tag can be “killed” permanently whenever an
item is purchased under endorsed RFID protocols [35].
2.6 The EPC Network
The Electronic Product Code (EPC) Network, originally developed by the Auto-ID
Center with its standards now managed by EPCglobal Inc., was designed and im-
plemented to enable all objects in the world to be linked via the Internet. The EPC
Network has undergone tremendous changes since it was first introduced. At the be-
ginning of this section, the early stage EPC Network will be introduced, followed by
the current state of an EPC Network.
A joint research paper on the EPC Network when it was first introduced had been
published [7, 8], which outlines components of EPC Network and their functionality.
2.6.1 The Early Stage EPC Network
Fig. 2.3 shows the structure of a typical EPC Network in the early stage. The EPC
Network consists of three major components, which are the Savant, the EPC Informa-
tion Service (EPCIS), and the Object Name Service (ONS). Strictly speaking, the reader
should be considered as part of the EPC Network. However, the reader is considered
to be a pure RFID tag interrogator under the control of a Savant, though implementa-
tions of some readers will integrate at least the base functionality of a Savant into the
reader itself.
• Savant
Page 19
2.6 The EPC Network
Figure 2.3. Early stage EPC Network.
The Savant is a middleware system located between a reader (or multiple read-
ers) and the applications in RFID systems. Applications operate on top of, or
within, the Savant operating environment. The Savant passes requests from the
application(s) to the reader(s) and receives unique tag identifiers and possibly
other data from sensors, and passes that information to the application(s). The Sa-
vant has several fundamental functions integrated into its design, some of which
are data filtering, aggregation and counting of tag data. These fundamental func-
tions are required in order to handle the extremely large quantity of data that
RFID systems can generate through the continuous interrogation of tags.
• Object Name Service
The function of the Object Name Service (ONS) in an EPC Network is to iden-
tify the location of the server hosting the appropriate information needed by an
application. In other words, the ONS acts like a “reverse phone directory” as
the ONS uses a number (EPC number) to retrieve the location (of data) from its
database. To encourage rapid development of the ONS, the ONS is purely based
on existing Internet technology and infrastructure. The first generation of ONS
system designs were based upon DNS systems with the first implementations
utilising existing DNS implementations with customised configurations.
• Physical Markup Language
Page 20
Chapter 2 RFID Systems
The Physical Markup Language (PML) defines the way information is transferred
in the EPC Network system. PML Core is based on the existing standard, XML
Schema Language. It uses “tags” to format the data before the data is sent. PML
does not use short “tags” for data formatting. Although more bandwidth will
be required for data transfer, compared to using short “tags”, the use of longer
descriptive tags increases human readability and hopefully will avoid mistakes
in interpreting and understanding the data, and how that data is to be handled.
• EPC Information Service
EPC Information Service (EPCIS) is the gateway between any requester of in-
formation and the database. It receives and sends messages in PML with any
requester of information, although it’s communication with the database can be
in any format or standard.
2.6.2 The EPC Network Current State
The new EPC Network is relatively different from the original version; it is more de-
fined and better structured. The explanation of the EPC Network at the current state in
this section is based on the EPCglobal Architecture Framework published in 2007 [41].
A whole picture of an EPC Network is as shown in Fig. 2.4.
Basically, there are two distinct types of components in the new EPC Network: soft-
ware/hardware role, and interface. They are self explanatory. Software/hardware role
components are either a piece of software or hardware which handle tasks allocated to
it. Interfaces govern how software/hardware role components communicate to each
other.
All the role components and interface components shown in Fig. 2.4 are explained
in detail in [41]. Some important parts are briefly highlighted to compare with the
original version of an EPC Network:
• Application Level Events
Page 21
2.6 The EPC Network
NOTE: This figure is included on page 22 of the print copy of the
thesis held in the University of Adelaide Library.
Figure 2.4. The Current EPC Network. Reproduced with permission from
[41] ©2006 EPCglobal.
Page 22
Chapter 2 RFID Systems
Ref. [42] is the complete specification of Application Level Events (ALE). It is an
interface component which handle the data flow of the filtering and collection on
the inventory of RFID tags.
• EPC Information Services
Ref. [43] is the complete specification of EPC Information Services (EPCIS). There
are several interface and role components bearing the name of EPCIS in Fig. 2.4.
Basically any component related to the EPCIS contributes to the data accessing,
sharing and storing between all the authorised RFID networks. By linking RFID
networks together, a global size network with real-time data sharing, also known
as an EPC Network, is made possible.
• Object Name Service
Ref. [44] is the complete specification of Object Name Service (ONS) as an inter-
face component. An ONS interface is used to locate the reference to the EPCIS
service or any other related other services. In Fig. 2.4, there are the Local ONS
and the ONS Root. They are related to but not the same as an ONS interface.
• Reader Management
Ref. [45] is the complete specification of Reader Management. As suggested by
its name, Reader Management is a role component which manages the operation
of RFID readers, mainly on the monitoring of the operating status and health of
RFID readers.
• Reader Protocol
Ref. [46] is the complete specification of Reader Protocol. Reader protocol is an
interface component. Basically, reader protocol allows data from RFID readers to
be passed up the EPCglobal Network stack, for example the ALE.
The basic idea of both the EPC Networks (early stage and current stage) is the same. It
is to enable real-time global data sharing for product tracking. The current EPC Net-
work is an updated version of the early stage EPC Network, retaining all the operating
Page 23
2.7 RFID Standards
principles while offering better clarity. The EPC Network is still evolving, depending
on the needs of RFID adopters.
2.7 RFID Standards
To enable and encourage wide spread adoption of RFID systems, one of the most im-
portant factors is to have RFID systems standardisation. Not only does this simplify
the hardware and software design for RFID systems, it also increases user confidence
in RFID.
Basically, there are two different areas of standardisation. The first one is to regulate the
spectrum usage and is always governed by local regulatory authorities. The common
standards are:
• “Title 47 part 15.247” published by Federal Communications Commission (FCC),
in 2001 [47]. This regulations is mainly adopted by the United States of America.
• “EN 302 208-1 v1.1.1: electromagnetic compatibility and radio spectrum matters
(ERM); radio frequency identification equipment operating in the band 865 MHz
to 868 MHz with power levels up to 2 W” published by European Telecommu-
nications Standards Institute (ETSI) for European Standard (Telecommunications
Series) in 2004 [48]. This regulation is mainly adopted by the countries in the
European Union.
• “AS/NZS 4771:2000 technical characteristics and test conditions for data trans-
mission equipment operating in the 900 MHz, 2.4 GHz and 5.8 GHz bands and
using spread spectrum modulation techniques, incorporating amendment No. 1”
published by Standards Australia.
The second type of standardisation is to have a uniform air interface and command
sets between an RFID reader and an RFID tag so that an RFID reader produced by a
company can be integrated into an RFID network setup by another company. There
Page 24
Chapter 2 RFID Systems
are two bodies responsible in providing standards, ISO (International Organization of
Standardization) [49] and EPCglobal [50].
The following standards are some of the standards available from ISO:
• For LF RFID operation: ISO 11784:1996: Radio-frequency identification of ani-
mals - code structure [51] and ISO 11785:1996: Radio-frequency identification of
animals - technical concept [52].
• For HF RFID operation: ISO/IEC 18000-3:2004: information technology - radio
frequency identification for item management - Part 3: parameters for air inter-
face communications at 13.56 MHz [33].
• For UHF RFID operation: ISO/IEC 18000-6:2004: information technology - radio
frequency identification for item management - Part 6: parameters for air inter-
face communications at 860 MHz to 960 MHz [53].
• For RFID terminology: ISO/IEC 19762-3, information technology - automatic
identification and data capture (AIDC) techniques - harmonized vocabulary -
part 3: radio frequency identification (RFID) [16].
The following standards are some of the standards available from EPCglobal:
• For the RFID operation in HF band, the standard is documented in “13.56 MHz
ISM band class 1 radio frequency identification tag interface specification: can-
didate recommendation, version 1.1.0” [54]. The HF Generation 2 Tag Proto-
col Standard is currently in development by the HAG HF Air Interface Working
Group.
• For the RFID operation in UHF band, the standard is documented in “EPC radio-
frequency identity protocols class-1 generation-2 UHF RFID protocol for com-
munications at 860 MHz - 960 MHz version 1.1.0” [35]. This version is consistent
with ISO/IEC 18000-6 Type C [53].
Page 25
2.7 RFID Standards
• As every tag contains a unique number and this unique number should be re-
trievable by any reader which conforms the the above stated standards, the tag
data contained by any tag must be standardised too. The standards is “EPC tag
data standards version 1.1 rev. 1.31” [4].
Page 26
Chapter 3
Path Loss and PositionAnalysis
Asimple path loss model for RFID systems is used to analyse
two collision problems in a dense reader environment. The
first problem is the reader interference with tag replies prob-
lem. The second problem is the effect of the “Listen Before Talk” (LBT)
regulations in European countries. It is shown that these two problems will
impede the wide-scale deployment of RFID systems. Based on a suggested
path loss model, a simulation script is written in MATLAB to visualise the
impact of reader interference with tag replies problem (the first problem)
and the effect of LBT regulations (the second problem) when multiple read-
ers are deployed in a same vicinity. With the help of this simulation soft-
ware, antenna positioning analysis was carried out efficiently.
Page 27
3.1 Introduction
3.1 Introduction
In a dense reader environment, there will be multiple RFID readers, and various types
of collision problems will reduce the reliability and efficiency of the RFID system.
RFID readers can use different channels to minimise collision. However, with limited
channels, collision will still occur in dense reader environments. The results can be
unsatisfactory with respect to read times and an unacceptable level of misreads may
result. In the worst case, collision problems could paralyse an entire RFID system.
Furthermore, collision problems also involve the disruption of the operation of other
RF devices in nearby frequency bands. Regulatory bodies have set some strict restric-
tions on RFID radiation to minimise this disruption problem. The concept of “Listen
Before Talk” (LBT) has been included in the European regulation, ETSI EN 302 208 [48],
which places severe restrictions on RFID deployment and causes some uncertainties
over the feasibility of RFID global deployment.
There are many types of collision problem and the following list shows some of the
more common types:
1. Reader Interference with Tag Replies (RITR), where transmissions from one reader
antenna interfere with tag reply signals in the receiver antenna of another reader.
2. Reader Induced Tag Confusion (RITC), where two reader antennas send simul-
taneous interrogating signals to a tag.
3. Regulatory Reader Shutdown (RRS), where a reader antenna is not allowed by
local regulations to transmit in a channel when there is a signal with a certain
strength detected in that channel.
The short-form names of these common collision problem, i.e. RITR, RITC or RRS, will
be used in this thesis. An alternate description of RFID collision problems can also be
found in [55].
There are two major objectives in this research:
Page 28
Chapter 3 Path Loss and Position Analysis
• To report on the analysis of the Reader Interference with Tag Replies (RITR) prob-
lem and provide a solution to RFID deployment regarding the RITR problem for
the benefit of those eager to set up RFID systems.
• To provide an analysis on the Regulatory Reader Shutdown (RRS) problem. The
focus will be on the restrictions imposed by “Listen Before Talk” to the placement
of RFID reader antenna.
The research performed to achieve these objectives is explained in the following se-
quence of elements:
1. A comprehensive background study on path loss models, with a decision made
later to use a piece-wise linear model for path loss in the RFID context:
Section 3.2 explains the theory of a path loss model, and shows how an in-building
path loss model is related to a free space model.
Section 3.3 explains the simple path loss experiments carried out around the lab-
oratory, and includes discussion of results.
Section 3.4 forms a path loss model, based on the results in Section 3.3, for the
RFID context.
2. A simple study of the Reader Interference with Tag Replies (RITR) problem based
on the chosen path loss model:
The path loss model chosen in Section 3.4 is used to investigate a simple two-
reader RITR situation in Section 3.5. Also, the safe distance between two RFID
reader antennas, before RITR occurs in a same channel, is computed and dis-
cussed.
3. Development of a simulation software based on the chosen path loss model:
Section 3.6 discusses the theory of power density, antenna gain pattern of an RFID
antenna, frequency channelling and the concept of “Listen Before Talk”. This
section is mostly informative but provides a strong background study for the
RFID positioning analysis discussion from Section 3.7 to Section 3.9.
Page 29
3.1 Introduction
Section 3.7 explains the concepts behind the simulation software developed in
MATLAB, and also, the input parameters for the simulation software to produce
the desired results.
4. Simulation results and discussions of the reader interference problem:
Section 3.8 re-performs the simple simple two-reader RITR in Section 3.5 using
the developed software from Section 3.7. The strength of this software over the
direct calculation method is examined.
5. Simulation results and discussions of the Regulatory Reader Shutdown (RRS)
problem:
Section 3.9 includes both the simulation results and discussions of various an-
tenna positioning strategies regarding the RRS problem. The focus will be on
the European “Listen Before Talk (LBT) provision. Intensive simulations were
carried out to visualise RFID reader signal strengths as received by isotropic or
directional receivers at various distances with respect to transmitting antennas,
either operating in the same channel or in various neighbouring channels.
6. Examples of some real life applications:
Some real life applications including the potential of reader synchronisation in
handling problems that may arise in a dense reader environment are presented
in Section 3.10.
It is hoped that this research can provide sufficient and useful guidelines on the safe
distance between antennas in a dense reader environment (to resolve the RITR prob-
lem) or even in a “Listen Before Talk” regulated area (to resolve the RRS problem). It
is also the belief of the author that by careful planning of antenna positioning, an RFID
system can be optimised and its interference to other RF systems in the surrounding
area can be minimised.
This chapter is the merging of two published papers [56, 57], with the study of reader
interference problem coming from [56] and the work on LBT analysis coming from [57].
Page 30
Chapter 3 Path Loss and Position Analysis
3.2 Path Loss Model
3.2.1 Free Space Path Loss
For a pair of lossless antennas in free space with optimum orientation we may write
the power transfer ratio in the form:
Pr
Pt= gtgr
(λ
4πd
)2
(3.1)
where λ = wavelength; Pt = transmitted power; Pr = available received power; gt =
transmitter antenna gain; gr = receiver antenna gain; d = separation distance between
antennas.
For some purposes it is desirable to separate the effects of antenna gain and distance
between antennas, and give the name free space path loss to the remaining factor in
the above equation. By expressing this factor in dB we have the free space path loss
expression:
PL(dB) = 20 log10
(4πd
λ
)(3.2)
Using the frequency of f = 915 MHz, λ = cf = 0.33 m, for a separation d = 1 m, PL(dB)
= 31.61 dB.
Also, (3.1) can be reexpressed as:
Pr(dBW) = Pt(dBW) + gt(dBi) + gr(dBi) − PL(dB) (3.3)
where Pt = transmitted power; Pr = available received power; gt = transmitter antenna
gain; gr = receiver antenna gain; PL = path loss.
3.2.2 In-building Path Loss
An in-building path loss is a path loss that occurs in a physical building, and it is
not the same as a free space path loss, as an in-building path loss will normally take
Page 31
3.3 Experiment
into consideration path obstruction, reflection, absorption and other attenuation effects
introduced by the presence of objects inside a building. Free space path loss has a
simple d−2 variation but in-building path loss has a more complex structure. The in-
building path loss model chosen for the purpose of simulation is given in (3.4), and is
explained by Rappaport [58]:
PL(dB) = PL(d0) + 10 × n × log10
(dd0
)(3.4)
where d0 = an arbitrary reference distance; n = a value that depends on the surround-
ings and building types; and d = the separation distance between the two antennas.
d0 will be chosen as 1 m and PL(d0), which is the in-building path loss at 1 m away, will
be approximately the same as the free space path loss at 1 m away, as room reflections
are not huge at this small distance. Hence, from (3.2), PL(1 m) is approximately 32 dB,
and (3.4) becomes:
PL(dB) = 32 + 10 × n × log10
(dd0
)(3.5)
As explained in (3.4), n is a empirical value that depends on the surroundings and
building types, and is only obtainable through experiment. An environment with high
n is a hostile environment for radiation, and its in-building path loss will be higher
when compared to the case of a low n environment.
3.3 Experiment
An interrogating RFID antenna was set to transmit a known signal while a measuring
spectrum analyser connected to a receiver antenna was moved away from the trans-
mitting antenna. The strength of the received signal was recorded versus the distance
away from the transmitting antenna. Removing antenna gain from the measured val-
ues gives us the values of path loss. The transmitting and receiving antenna used in
this experiment both have a gain of 6 dBi. Fig. 3.1 is plotted using logarithmic scale,
and we have a straight line approximation of:
Page 32
Chapter 3 Path Loss and Position Analysis
PL(dB) = 31 + 40 log10
(d1
)(3.6)
for d > 8 m. The results resemble the model based on (3.7) in Section 3.4 and also
strengthen the belief that reader interference problem must be solved for large scale
RFID deployment.
0
20
40
60
80
100
120
1 10 100
Distance of Separation Between Antennas, d (m)
Pat
h L
oss
(dB
)
Figure 3.1. Experimental path loss results. Computed path loss results from experiment, with
variation in the distance between the transmitting and receiving antennas.
The in-depth details of the experiments, including equipment setup, experiment venues
and data analysis can be found in Section 4.4 and Appendix F.
3.4 Path Loss Model for RFID
For the theoretical calculations in this chapter, we chose n = n2 = 3.5, when the separa-
tion is more than or equal to 8 m, and we chose n = n1 = 2.5 when the distance is less
than 8 m. i.e. we are using approximation equations:
PL(dB) =
⎧⎪⎪⎪⎨⎪⎪⎪⎩
PL(d0) + 10 × n1 × log10dd0
0 < d ≤ 8 m
C + 10 × n2 × log10dd0
d ≥ 8 m
(3.7)
Page 33
3.5 Simple Two Reader Interference
where PL(d0) is 31.61 dB at 915 MHz (from (3.2)). For all the analysis presented in this
thesis, PL(d0) takes the approximate value of 32 dB. C is constant value with respect
to the chosen PL(d0), n1 and n2 values to ensure continuity of (3.7) at position d = 8 m.
For a PL(d0) of 32 dB, n1 of 2.5 and n2 of 3.5, C takes the value of 22.97 dB.
Hence, effectively, the in-building path loss model from (3.5) had been modified to the
piece-wise linear (in log10dd0
) relation given in (3.7) and plotted as shown by the heavy
line in Fig. 3.2.
0
20
40
60
80
100
120
140
1 10 100
Transmitter-Receiver Separation (m)
Pa
th L
oss (
dB
)
n=4
n=3
n=2
n=6n=5
Figure 3.2. Plot of proposed piece-wise linear in-building path loss model against distance.
Dotted lines represent the path loss model with environment factor, n, from 2 to 6
respectively, calculated using (3.5). The bold line represents the suggested piece-wise
linear model, represented by (3.7).
3.5 Simple Two Reader Interference
As highlighted in the introduction section of this chapter, one type of reader collision
chosen for study is the Reader Interference with Tag Replies (RITR) problem. RITR
happens when transmitting signals from a reader antenna drown reply signals of a tag
to another reader antenna.
In this section, a scaled down version of RITR is discussed, where the RITR involves
only two readers, and the readers are assumed to transmit and receive in the same
Page 34
Chapter 3 Path Loss and Position Analysis
channel of a multi-channel frequency band. Consider the case where there are two
readers, A and B, using the same channel, Channel C. It is assumed that Reader A and
Reader B are identical, the antennas for both of them are the same and have the same
gain. Also, both of the antennas are facing each other. Reader A uses Channel C to
interrogate a tag (which is located directly in front of antenna of Reader A), and the
tag will have in-band backscattering in response to Reader A. If the power delivered
to both reader antennas is 0 dBW, with reader antenna gains of 6 dBi, a total of 6 dBW
EIRP will be obtained. Fig. 3.3 shows diagrammatically how the interrogation between
Reader A and the tag occurs. Signal paths are shown using the bold dotted lines. At
1 m away, the path loss for signals 1 and 2 obtained using (3.7) is 32 dB. Hence the total
path loss is approximately 64 dB (2 × 32 dB).
Reader A Reader BTag
3
1
2
Reader
Antenna
Figure 3.3. A simple illustration of Reader Interference with Tag Replies (RITR). RITR will
occur when a reply from a tag is interfered with by a signal sent from another nearby
reader. In the example illustrated, tag reply 2 is interfered with by interrogation signal
3 from Reader B. The bold dotted line (signal path) signifies that the two reader
antennas are facing each other, while the tag is in-line with the maximum gain direction
of antenna A. The numbered arrows indicate the distances and directions of signal
propagation between objects of interest. The numbered arrows do not represent the
actual propagation paths.
The tag antenna has a gain of approximately 1.5. However, to take in consideration
a probable tag polarisation misalignment, i.e. the tag antenna is not oriented in the
most optimal way, it is assumed that the tag has a unity antenna gain. Also assumed,
is that the efficiency of the tag is 10%, and the signal will suffer another 10 dB loss.
All the losses (path loss + tag efficiency-tag antenna gain) are summed up to be 74 dB.
Since Reader B is also using the same channel, Channel C, the interrogation signal sent
by Reader B will interfere with the in-channel backscattered signal from the tag. The
Page 35
3.5 Simple Two Reader Interference
Table 3.1. The effect of tag distance on multi-reader interference. Referring to Fig. 3.3,
Dtag−A is the distance between the tag and reader antenna A; MinB int is the minimum
distance between the reader antenna A and reader antenna B, before the signal from
reader antenna B interferes with the tag reply to antenna A; Df s is MinB int in free
space.
Dtag−A (m) Total Loss (dB) MinB int (m) Df s (m)
1 74 28.7 130.7
2 89 76.9 735.2
5 109 286.5 7,352.5
10 123 718.3 35,237.1
question is how near Reader B needs to be to interfere with the backscattered signal.
The comparison is made between (a) path loss of path 1 and path 2, tag antenna gain,
and tag efficiency loss, and (b) path loss of path 3. This comparison is only applicable
in the situation as described in this section. In this section, the losses in both (a) and (b)
are called Total Loss. Of course in the usual scenario, the values are different for the
two cases.
Note that the gains of reader antennas (for both Reader A and B) should be considered
as well. However, since they contribute a total of 12 dB to both case (a) and case (b),
they are not included in the comparison calculations.
Path 3 in Fig. 3.3 shows the distance and direction of signal travelling along the dotted
line from Reader B to Reader A. Using (3.7) to calculate path loss, a distance of around
28.7 m is needed to have path loss, or total loss in this case, of 74 dB. Table 3.1 shows
some results on the minimum distance for Reader B to interfere with the tag reply.
Again, the results in columns 2 and 3 of Table 3.1 are computed using (3.7), and (3.7)
takes into consideration in-building propagation loss. It is very natural to also raise
the question of what will be the case, if free space propagation loss is considered. The
results obtained using free space propagation loss model are attached as column 4 in
Table 3.1.
Page 36
Chapter 3 Path Loss and Position Analysis
It is discovered that, if a tag is located 10 m away from an interrogating reader, the
antenna of the interfering reader must be around 718.3 m away. Since a state of art
reader in the market can have a read range of around 10 m when reading a passive tag,
to put the next reader several hundred metres away is not sensible. Section 3.10 in the
latter part of this chapter provides ideas to solve this problem.
We continue to make the assumption that free space path loss is not applicable in prac-
tical applications, and that at the distances emerging from these calculations, at least
some obstacles will be present, and a within-building propagation loss model is ap-
propriate. As shown in Table 3.1, when free space path loss model is used, the safe
distance between two antennas facing each other to prevent interference becomes rel-
atively huge (> 35 km), but we do not believe this model is relevant to practical situa-
tions.
3.6 Background for RFID Positioning Analysis
This section presents important background studies which assist the detailed discus-
sion of RFID positioning analysis. Firstly the difference between EIRP (Equivalent
Isotropic Radiated Power) and ERP (Effective Radiated Power) is discussed, followed
by the examination of a frequency spectrum channelling technique used in RFID oper-
ation. Also included is the in-depth investigation of the “Listen Before Talk” provision
in the European RFID regulation.
3.6.1 Power Density
Gain of an antenna is the ratio of the radiation intensity, in a given direction, to the
radiation intensity that would be obtained if the power accepted by the antenna were
radiated isotropically.
Power density, S, which at a distance d for a transmitting antenna in the direction of
maximum gain, has the value:
Page 37
3.6 Background for RFID Positioning Analysis
S =(
EIRP4πd2
)(3.8)
where Equivalent Isotropic Radiated Power (EIRP) is defined as, in a given direction,
the gain of a transmitting antenna multiplied by the net power accepted by the antenna
from the connected transmitter. In simple term, EIRP is the product of power delivered
to the antenna (Pt) and antenna gain with respect to the gain of an isotropic radiator
(gt):
EIRP = gtPt (3.9)
Effective Radiated Power (ERP) is a more popular term in Europe and is defined as,
in a given direction, the relative gain of a transmitting antenna with respect to the
maximum directivity of a half-wave length dipole (or alternatively the maximum gain
of a lossless half-wave length dipole) multiplied by the net power accepted by the
antenna from the connected transmitter. In simple term, ERP is the product of power
delivered to the antenna (Pt) and antenna gain with respect to the maximum gain of a
lossless half-wave length dipole (gd):
ERP = gdPt (3.10)
It should be noted that the “E” in EIRP stands for equivalent while the “E” in ERP
stands for effective. Since the maximum gain of a lossless half wave length dipole
relative to an isotropic radiator is 1.64, gd can be expressed in term of gt:
gd =gt
1.64(3.11)
With (3.11), ERP can be expressed in term of gt:
ERP =gtPt
1.64(3.12)
Thus, by dividing (3.9) and (3.12), and rearranging:
Page 38
Chapter 3 Path Loss and Position Analysis
ERP = 1.64 × EIRP (3.13)
Hence, (3.8) can be reexpressed as:
S =(
1.64 × ERP4πd2
)(3.14)
3.6.2 Antenna Gain Pattern
As shown by (3.1), antenna gain must be taken into consideration, as well as path loss,
to relate the power delivered to a transmitting antenna to power received by another
antenna. A typical RFID antenna is a directional circularly polarised antenna with a
gain of 6 dBi. The polar plot of the gain of a typical RFID antenna is shown in Fig. 3.4,
and has been used, when directional antennas are called for, to simulate and analyse
the antenna positioning in the dense reader environment presented in this thesis.
30
210
60
240
90
270
120
300
150
330
180 0
6 dBi 10
0
−10
−20
Figure 3.4. Polar plot of the antenna gain of a directional circularly polarised RFID antenna
with a gain of 6 dBi. This plot is not a direct measurement plot. An anechoic chamber
measurement was carried out on a 6 dBi RFID antenna. The measurement results are
then sampled at discrete points before input into MATLAB for plotting.
Page 39
3.6 Background for RFID Positioning Analysis
3.6.3 Frequency Spectrum Channelling
Frequency spectrum channelling is a popular technique used in reducing both types
of reader collision problem considered in this chapter. One of the biggest challenges
in frequency spectrum channelling is the sideband interference, where the sideband
of a particular transmitting channel interferes with signals in a neighbouring channel.
Under the latest EPCglobal UHF Class 1 Generation 2 protocol [35], the sidebands of
a transmission must adhere to the limits as shown in Fig. 3.5. The dBch is defined as
decibels referenced to the integrated power in the reference channel. The simulation
software developed also uses this limit to investigate interactions of readers operating
using different frequency channels. This investigation is presented in detail in Sec-
tion 3.9.4.
There are two different types of transmit mask, one for the so-called multiple-interrogator
environment and the other for the so-called dense-interrogator environment. Accord-
ing to [35], the multiple-interrogator environment is defined as an operating environ-
ment (in free space, a sphere with radius of approximately 1000 m) within which a
modest number of the available channels are occupied by active interrogators (for ex-
ample, 5 active interrogators operating in 25 available channels). The dense-interrogator
environment is defined as as an operating environment (in free space, a sphere with
radius of approximately 1000 m) within which most or all of the available channels are
occupied by active interrogators (for example, 25 active interrogators operating in 25
available channels).
Note that in explaining the difference between a multiple-interrogator environment
and a dense-interrogator environment, [35] defines a sphere with radius of approxi-
mately 1000 m in free space as the operating environment. Referring to (3.7), the path
loss at 1000 m away using n1=n2=2 (free space) is approximately the same as the path
loss at 100 m away when n1=2.5 and n2=3.5 (the suggested path loss model for RFID).
Hence the operating environment can be restated as a sphere with radius of approxi-
mately 100 m in real life.
As stated before, more details on frequency spectrum channelling studies will be pro-
vided in Section 3.9.4.
Page 40
Chapter 3 Path Loss and Position Analysis
NOTE: This figure is included on page 41 of the print copy of the
thesis held in the University of Adelaide Library.
Figure 3.5. Transmit mask for multiple-interrogator and dense-interrogator
environments.
Reproduced with permission from [35] ©2006 EPCglobal. dBch is defined as the
decibels referenced to the integrated power in the reference channel. Tari is the
reference time interval for a data-0 in Interrogator-to-Tag signalling, and is between
6.25 µs and 25 µs. fc is the centre frequency of a transmission.
3.6.4 Listen Before Talk
ETSI 302 208 is a European regulation governing the operation of RFID readers [48].
It allocates the frequency band of 865 to 868 MHz for RFID deployment. This
frequency band is then divided into fifteen sub-bands or channels, each spanning a
total of 200 kHz. However, when a reader is operating at the maximum radiated
power, which is 2 W ERP (Effective Radiated Power), only ten sub-bands are
available, while the remaining five are utilised as guard bands or for lower power
readers. ETSI 302 208
Page 41
3.6 Background for RFID Positioning Analysis
also introduces the concept of “Listen Before Talk”. An extract from the ETSI 302 208
best describes the essence of “Listen Before Talk”. It states “Prior to each transmission,
the receiver in the interrogator shall first monitor in accordance with the defined listen
time for the presence of another signal within its intended sub-band of transmission.
The listen time shall comprise a fixed period of 5 ms plus a random time of 0 ms to
5 ms in 11 steps. If the sub-band is free the random time shall be set to 0 ms” [48]. The
threshold to determine the presence of another signal within the intended sub-band is
shown in Table 3.2. The measurement method is defined in the standard.
Furthermore, once a sub-band has been selected, the RFID reader is permitted to use
that sub-band for up to 4 s. After use, it must free the sub-band for at least 100 ms.
A reader can however, listen to another sub-band for 5 ms and if free use that new
sub-band immediately.
Table 3.2. UHF RFID reader radiated power with corresponding threshold values for LBT,
and minimum allowable distance between antennas. Threshold value obtained from
[48]. Minimum path loss and distance required between a transmitting and a receiving
antennas are calculated with the assumption that both antennas are isotropic radiators.
ERP (W) ERP (dBW) Threshold (dBW) Path Loss (dB) Distance (m)
Up to 0.1 Up to -10 ≤-113 103 193.5
0.1 to 0.5 -10 to -3 ≤-120 117 485.9
0.5 to 2.0 -3 to 3 ≤-126 129 1070.1
Similarly to the results shown in column 2 and 3 of Table 3.1 in Section 3.5, the “Dis-
tance” column in Table 3.2 is computed using (3.7), where we consider an in-building
propagation model with n value set to 2.5 for distance less than 8 m and n value set
to 3.5 for distance more than 8 m. As mentioned before, the value of n changes from
building to building. The path loss model chosen for RFID operation is just a good
approximation.
The main point here is that if RFID readers are going to be deployed on a large scale,
most likely the system will not work in optimal operation mode as the Regulatory
Page 42
Chapter 3 Path Loss and Position Analysis
Reader Shutdown (RRS) problem will occur. This is due to the fact that LBT will effec-
tively shut down many of the channels, though those channels might have been, in the
absence of the LBT provision, freely available for interrogation between readers and
tags. This RRS problem is one of the collision problem in RFID deployment area which
is covered in this chapter. More details on the study of RRS problem will be provided
in Section 3.8.
3.7 Simulation Concepts
Section 3.5 shows that a path loss model can be used to analyse the Reader Interference
with Tag Replies (RITR) problem. However, as explained in Fig. 3.3, the two reader
antennas are positioned face-to-face and the maximum gain is used in both paths of
the calculation and in consequence has no effect. If the reader antennas are of different
orientations, the calculation will be very tedious and the antenna gain pattern has to
be referred to all the time. Hence, a general purpose simulation software was designed
to handle more complex situations, with advanced features that include:
• Handling of more than two reader antennas
Theoretically, the software can simulate an unlimited number and variety of an-
tennas. However, the computational time will increase linearly with linear incre-
ment of the numbers of reader antennas.
• Full consideration of the antenna gain pattern
The antenna gain pattern for every type of antenna needs to be imported to the
simulation software, but only once. After that, the simulation software can con-
sider the gain pattern for any antenna configuration.
• Results in graphical form
The simulation results will be presented graphically. This makes the interpreta-
tion of the results easy.
The simulation software is used to contribute to the reduction of two distinct collision
problems in dense RFID reader environment:
Page 43
3.7 Simulation Concepts
1. The Reader Interference with Tag Replies (RITR) problem.
Similarly to Section 3.5, the simulation software can compute the safe distance
between reader antennas to avoid the RITR problem.
2. The Regulatory Reader Shutdown (RRS) problem in European countries.
As discussed in Section 3.6.4, European countries are subjected to the provision
of “Listen Before Talk” (LBT). The simulation software is used to demonstrate the
constraints in deploying RFID reader antennas introduced by LBT.
The simulation script used is written in MATLAB code to show the power received
by an isotropic antenna (0 dB gain) at any location. The simulation concepts can best
be described with the help of Fig. 3.6. Firstly, a grid with size defined by the user is
formed. The centre of each grid box (not the intersection points of grid lines) can be
used as the location of one or more transmitting reader antennas or as the location of
an isotropic receiving antenna. In normal simulation, only the positions of transmit-
ting antennas need to be specified, while all other unassigned grid boxes will become
isotropic receiving antennas. For example in Fig. 3.6, a transmitting antenna is located
at position (4,1) and is symbolised by × and the rest of grid boxes become (unless
otherwise stated) isotropic receiving antennas, which are symbolised by •.
The resolution setting in the simulation script defines the distance between two neigh-
bouring isotropic receiving antennas (distance between the centres of two horizontally
or vertically neighbouring grid boxes). By default the resolution is set to be 1 m. An-
other common resolution used is 10 m. The resolution can also be set lower than 1 m.
The choice of resolution depends on the accuracy requirement for results and also de-
pends how wide is the solution space. For example, it is not wise to have resolution of
0.01 m and the solution space of 1000 m × 1000 m.
All the transmitting antennas used in the simulation are directional circularly polarised
antennas (unless otherwise stated), with antenna gain pattern shown in Fig. 3.4, and
are located 1 m above ground level (refer to Fig. 3.7 for visualisation of the transmitting
antenna configuration). Hence the grid shown in Fig. 3.6 lies in the x-y plane shown in
Fig. 3.7.
Page 44
Chapter 3 Path Loss and Position Analysis
Extended
Grid
Extended
Grid
Figure 3.6. Grid used in simulation software. The symbol × represents a possible location of a
reader antenna while the symbols • indicates the positions where the received powers
of an isotropic receiver antenna are simulated.
The software uses the path loss model as shown in (3.7). By specifying the power
delivered to the transmitting antenna, (3.3) is used to compute the received power with
the gain specified in Fig. 3.4, path loss from (3.7), and a unity gain for the receiving
antenna.
The simulation result (example Fig. 3.8) is in graphical form and is able to show the
power received by an isotropic antenna (0 dB gain) at any location, bounded by the
solution space set by the user. A broader solution space means more simulated points
and requires longer simulation time.
The orientation and the position of a transmitting antenna with respect to the simu-
lation grid can be illustrated in Fig. 3.7. The simulation results represent the signal
strength received by an isotropic antenna in the x-y plane.
In summary, the developed software can take in the following input:
Page 45
3.7 Simulation Concepts
1 m
Default AntennaConfiguration
Ground
z
yx
Figure 3.7. The orientation and the position of a transmitting antenna with respect to the
simulation grid. The simulation software computes the signal strength at the x-y plane.
By default the x-y plane is 1 m above the ground plane. In Section 3.9.3, the software
is modified so that the elevation of the x-y plane can be changed to any height with
respect to the ground level.
• Position of transmitting antennas: The position of each individual transmitting
antenna is input as cartesian coordinates (x,y). In theory there is no upper limit
on the number of antennas which can be included in the simulation. Nonetheless
the computation time increases in a linear fashion with the increment of the total
number of transmitting antennas.
• Radiation pattern: The radiation pattern of each individual transmitting antenna
is input as a set of discrete values. Unless otherwise stated, for all the simulations
presented in this thesis, the radiation pattern as shown in Fig. 3.4 is used.
• Power delivered to the transmitting antenna: This the power input into the reader
antenna. By default, 1 W is used.
• Solution space: This parameter determines the simulation area. For example, the
area of simulation can be set to 10 m × 10 m.
• Resolution: Resolution setting controls how frequent (with respect to distance)
the simulation software computes the signal strength within the solution space.
For example, the software can calculate every 0.1 m or every 1 m.
Page 46
Chapter 3 Path Loss and Position Analysis
3.7.1 Simulation Concepts Example
5 10 15 20 25 30 35 40 45 50
5
10
15
20
25
30
35
40
45
50−64.5
−61
−57.5
−54
−50.5
−200
−180
−160
−140
−120
−100
−80
−60
−40
−20
0
Around−52 dBW
Figure 3.8. Simulation results verifying the functionality of the software developed in MAT-
LAB. Results from simulation for simulation concepts example (vertical and horizontal
axes in m, representing the y and x axes shown in Fig. 3.7, received power in dBW). The
simulated antenna is located at position (25,25) facing in the y direction. At a distance
of 10 m away, the signal strength is approximately -52 dBW, which is as predicted by
the path loss model.
Fig. 3.8 shows a simple example of simulation results. In this example, the default
value of 0 dBW of power is fed through a 6 dBi gain antenna of pattern shown in
Fig. 3.4, resulting in a 4 W EIRP radiation. The 6 dBi gain is only valid in the front
direction of the antenna. Hence, at a distance of 10 m away directly in front of the
antenna, the path loss will be 58 dB, as computed from (3.7). The received power at that
location as detected by a 0 dB gain antenna is hence 0 dBW (transmit power) + 6 dBi
(transmitter antenna gain) -58 dB (path loss) + 0 dBi (receiver antenna gain), giving a
predicted received power of -52 dBW. In Fig. 3.8, a horizontal guideline is added to
show the point located 10 m away from the antenna. At that point, the received power
is calculated by the simulation software to be approximately -52 dBW, which agrees
with the more directly calculated result.
Page 47
3.8 Results and Discussion (RITR)
3.8 Results and Discussion (RITR)
In this section, the results obtained in Section 3.5 (using direct calculation) are repro-
duced using a modified form of the simulation software described in Section 3.7. The
discussion also shows the strength of the developed software over the direct calcula-
tion method. The usage of this same software in predicting Regulatory Reader Shut-
down (RRS) problem is presented in next section. This section only focuses on the
Reader Interference with Tag Replies (RITR) problem.
3.8.1 Modification of Existing Software
As the simulation software described in Section 3.7 is in a very general form to cater
for various applications, a few modifications were carried out to compute the effect of
tag distance on the RITR problem. The normal input parameters (from Section 3.7) are
still required. The additional inputs required are:
• P1(x,y,d): Position and direction of a first transmitting reader antenna (represent-
ing antenna A in Fig. 3.3). d is the direction indicator of the antenna. Currently
the antenna can either point in the x or in the y direction (Refer Fig. 3.7). This is
the antenna experiencing interference problem.
• T(x,y): Position of the tag of interest (representing the tag of interest in Fig. 3.3).
It is assumed that this tag has unity gain in all directions. This tag is sending
replies to the first transmitting reader antenna.
• P2(x,y,d): Position and direction of a second transmitting reader antenna (repre-
senting antenna B in Fig. 3.3) d is the direction indicator of the antenna. Cur-
rently the antenna can either point in the x or in the y direction (Refer Fig. 3.7).
This is the antenna producing interference signals to the first transmitting reader
antenna.
With the above input values, the modified software will execute the following steps:
Page 48
Chapter 3 Path Loss and Position Analysis
1. Run a simulation with the specified solution space, specified resolution, with the
first transmitting antenna located at position P1(x,y) and pointing direction P1(d),
with default gain pattern (Fig. 3.4).
2. The signal strength as received by an isotropic receiver at position T(x,y) is recorded.
As it is assumed that tag has an efficiency of approximately 10%, this recorded
signal strength is further reduced by 10 dB, and is stored as temptag.
3. Run a second simulation with the specified solution space, specified resolution,
with the first transmitting antenna removed, but with an isotropic radiator lo-
cated at T(x,y) and power delivered to this isotropic radiator set to be temptag.
The signal strength received by an isotropic receiver at position P1(x,y) is recorded
as Stag. The angle between the tag position T(x,y) and the position of the first
transmitting antenna P1(x,y) is then computed to obtain the right gain value for
the first transmitting antenna. This gain is used to increase the Stag value. This
Stag represents the strength of tag replies as received by first transmitting an-
tenna. Note that first transmitting antenna does not appear in this second run of
simulation but the position of this first transmitting antenna, P1(x,y), is required.
4. Run a third simulation with the specified solution space, specified resolution,
with only the second transmitting antenna located at position P2(x,y) and point-
ing direction P2(d), with default gain pattern (Fig. 3.4). The signal strength re-
ceived by an isotropic receiver at position P1(x,y) is recorded as Sreader. The angle
between the second transmitting antenna and the first transmitting antenna is
then computed to obtain the right gain value for the first transmitting antenna.
This gain is used to increase the Sreader value. This Sreader represents the strength
of signals from the second transmitting antenna as received by the first transmit-
ting antenna.
5. Stag and Sreader will be the output values. When Sreader > Stag, there will be a
Reader Interference with Tag Replies (RITR) problem.
It can be seen that antenna gain pattern is referred to frequently in the calculation,
especially when the two transmitting antennas and tag of interest are not aligned in a
Page 49
3.8 Results and Discussion (RITR)
straight line, as they were in Fig. 3.3. This software by taking into account the antenna
gain pattern and can give a quick indication whether a RITR problem will happen or
not based on the position of the two transmitting antennas and the position of the tag
of interest.
3.8.2 Guidelines to Avoid RITR
In common practice, the relative position and direction of two RFID reader antennas
is in one of the forms shown in Fig. 3.9. Hence it is the aim of this section to provide a
safe distance guideline for antennas in the situations shown in Fig. 3.9 to avoid RITR.
Referring to the antenna gain pattern shown in Fig. 3.4, the gain of the antenna in
front, behind and at the sides are, 6 dBi, -22 dBi, and -14 dBi respectively. The safe
distances for these configurations are obtained through simulations and are presented
in Table 3.3.
The safe distance between two antennas with a angle between them (such as the one
shown in Fig. 3.9 (h), depends heavily on the radiation pattern, the direction of maxi-
mum radiation, and the angle, θ. Hence, the safe distance cannot be presented here as
a guideline but can only be obtained on a case by case basis through simulation.
By following the safe distance guideline tabulated in Table 3.3, RITR can be reduced
but is not avoided. For example, if two 6 dBi directional antennas are to be positioned
back to back (Fig. 3.9 (d)), and tags are to be read not more than 1 m away, the safe
distance between these two antennas is 0.3 m. From the same table, it can be noticed
that it is not advisable to design an RFID system with tags to be read at a far distance
away, as an increased distance between a tag and a reader antenna interrogating the
tag will result in an increment in the safe distance between reader antennas to prevent
the RITR problem.
Page 50
Chapter 3 Path Loss and Position Analysis
(a) (b) (c)
(e) (f)
(g) (h)
(d)
Figure 3.9. Common positionings of antennas with respect to each other. The coloured
antenna is referred as the first antenna. (a) antennas facing each other; (b) antennas
side-by-side facing in the same direction; (c) antennas side-by-side but facing in opposite
directions; (d) antennas back-to-back; (e) antennas in a line facing in the same direction;
(f) second antenna facing the side of the first antenna; (g) second antenna backing the
side of the first antenna; (h) other positions with an angle between antennas, and
radiation directions as shown.
Page 51
3.9 Results and Discussion (RRS)
Table 3.3. Safe distances, in m, for different antenna positioning to avoid the Reader In-
terference with Tag Replies (RITR) problem. The different antenna positionings are
defined in Fig. 3.9. All these results are obtained through simulations using the simu-
lation script described in Section 3.7. Note that these results are for cases where both
the transmitting and receiving antenna are directional and with antenna gain pattern
shown in Fig. 3.4. Dtag−A is the distance between the tag and reader antenna A shown
in Fig. 3.3.
PositioningDtag−A
1 (m) 2 (m) 5 (m) 10 (m)
Fig. 3.9 (a) 28.7 77.3 286.2 718.3
Fig. 3.9 (b) 1.2 4.8 20.6 51.7
Fig. 3.9 (c) 1.2 4.8 20.6 51.7
Fig. 3.9 (d) 0.3 1.1 6.9 18.0
Fig. 3.9 (e) 3.6 12.2 45.4 113.8
Fig. 3.9 (f) 7.6 20.7 76.8 192.7
Fig. 3.9 (g) 0.6 2.3 12.2 30.5
3.9 Results and Discussion (RRS)
The following simulations are completed following the concepts outlined in Section 3.7.
The path loss model used is shown in (3.7), with n = 2.5 for near distance and n = 3.5
for far distance. Also, the focus will be on the effect of LBT on the deployment of RFID
reader antenna which results in the Regulatory Reader Shutdown (RRS) problem. In
most of the simulation results, the -126 dBW (the LBT threshold for 2 W ERP operation
shown in Table 3.2) boundary is emboldened.
Page 52
Chapter 3 Path Loss and Position Analysis
−91.1
−86.1
−96.1 −101 −106 −111 −116 −121
−126
Figure 3.10. Simulation results on one antenna simulation. Results from simulation (vertical
and horizontal axes in m, representing the y and x axes shown in Fig. 3.7, received power
in dBW). For a second antenna with the same y coordinate to utilise the same channel
for transmission while adhering to the LBT provision, the next possible isotropic an-
tenna must be 350 m away. If using the same y coordinate but using an antenna with
side gain of approximate -14 dB, this antenna must be approximately 150 m away.
3.9.1 One Antenna Simulation
The one antenna simulation described here is similar to the example given in Sec-
tion 3.7.1. The only difference is the simulation here covers a larger area to identify
available zones to deploy a second antenna working in a same channel. From Table 3.2,
the maximum radiated power allowed by current European regulations is 2 W ERP, or
equivalently 3.2 W EIRP. From Table 3.2, the minimum threshold is -126 dBW. The
simulation result is shown in Fig. 3.10, and shows that for the next antenna to be given
the same y coordinate and to be able to operate in the same channel (assuming the
second antenna is an isotropic radiator) with the initial antenna under consideration,
the antennas must be separated by a distance of around 350 m away in the horizontal
direction.
If the second antenna (still an isotropic radiator) is to be placed in the rear position
of the first antenna, the minimum distance between them would be shorter, and is
Page 53
3.9 Results and Discussion (RRS)
Table 3.4. Safe distance for different antenna positionings to avoid Regulatory Reader Shut-
down (RRS) problem. The different antenna positionings are defined in Fig. 3.9. All
these results are obtained through simulations using the simulation script described in
Section 3.7. Note that these results are for cases where both the transmitting and
receiving antenna are directional and with antenna gain pattern shown in Fig. 3.4.
Positioning Safe Distance (m)
Fig. 3.9 (a) 1900
Fig. 3.9 (b) 150
Fig. 3.9 (c) 150
Fig. 3.9 (d) 50
Fig. 3.9 (e) 300
Fig. 3.9 (f) 500
Fig. 3.9 (g) 80
approximately 210 m. This result is not obtainable from Fig. 3.10, but is from another
run of simulation, where the transmitting antenna is positioned is a suitable position
to give more view on the area behind the transmitting antenna.
However in practice, most likely the second antenna is not an isotropic radiator, and
will be the same model as the first antenna with a gain of 6 dBi. Hence the orientation
of the second antenna will be crucial to determine the minimum distance between the
first and second antennas.
A similar analysis to the analysis performed in Section 3.8.2 was carried out. Again,
referring to the antenna gain pattern shown in Fig. 3.4, the gain of the antenna in front,
behind and at the sides are, 6 dBi, -22 dBi, and -14 dBi respectively.
This gain must be considered when deciding where the second antenna can be posi-
tioned. Common positioning configurations have been shown in Fig. 3.9. The safe
distances for these configurations are obtained through simulations and are presented
in Table 3.4.
Page 54
Chapter 3 Path Loss and Position Analysis
Also, similarly to the analysis presented in Section 3.8.2, the safe distance between two
antennas with a angle between them (such as the one shown in Fig. 3.9 (h), depends
heavily on the radiation pattern, the direction of maximum radiation, and the angle,
θ. Hence, the safe distance cannot be presented here as a guideline but can be only be
obtained on a case by case basis through simulation.
3.9.2 Two Antenna Simulation
From the one antenna simulation, it has been shown that another antenna operating
in the same channel can be deployed around 350 m horizontally away from the first
antenna if the second antenna is a isotropic radiator. In practice, where a directional
6 dBi antenna is used (Fig. 3.4), the second antenna can be placed 150 m horizontally
away from the first antenna, as is indicated in Table 3.4.
In the two antenna simulation discussed here, a second antenna is deployed 150 m
horizontally away from the first one, and the result is as shown in Fig. 3.11. The bold
line is the approximate safe boundary for the deployment of an isotropic radiator in the
same channel. The distance is approaching 1.6 km in front of both antennas and around
400 m horizontally away from the second antenna. If a directional 6 dBi antenna is
used as the third antenna (which is most likely the case in reality), the safe distance
depends on the positioning and orientation of the third antenna. If the placement of
third antenna takes the form of Fig. 3.9 (a), facing directly to either of the first two
antenna, it has be to placed approximately 2.2 km away (outside the solution space
shown in Fig. 3.11). If the third antenna is to be positioned side-by-side with the first
two antennas, the safe distance will be around 150 m away from the nearest antenna
(well within the solution space shown in Fig. 3.11).
This shows that for antennas transmitting at the same frequency even with care in
orientation, they have to be at a great distance apart to avoid RRS problem and hence
this deployment is not practical in real life applications where LBT is enforced.
Page 55
3.9 Results and Discussion (RRS)
−129
−126
−124
−121
Figure 3.11. Simulation results of a two antenna simulation. Results from simulation (vertical
and horizontal axes in units of 10 m, representing the y and x axes shown in Fig. 3.7,
received power in dBW). The two antennas are positioned at locations (50,50) and
(200,50) respectively. The bold line is the approximate safe boundary for the deploy-
ment of an isotropic radiator in the same channel.
3.9.3 One Antenna in Horizontal Position
Another antenna placement configuration is to place an antenna in a horizontal po-
sition, with the direction of maximum gain perpendicular to the ground plane. An
antenna in this fashion can have the interrogation field projected upwards or down-
wards, as shown in Fig. 3.12. Whenever the term horizontal antenna is mentioned
throughout this chapter, it is referred to the configuration as illustrated in Fig. 3.12
with a certain elevation, which will need to be specified either as a default value of 0 m
or as a separately specified value that overrides the default.
An antenna in a horizontal position can be on any elevation level with respect to the
ground. A UHF reader antenna is usually a patch antenna and the patch will be parallel
with the ground and may be radiating towards the ground. However, the developed
simulation software does not consider the reflection from the ground. Hence, the de-
veloped software is more accurate in the case where there is no antenna facing towards
ground.
Page 56
Chapter 3 Path Loss and Position Analysis
1 m
1 m
Default AntennaConfiguration
Ground
Horizontal Antenna(0 m Elevation, Upwards)
Horizontal Antenna(2 m Elevation, Downwards)
Figure 3.12. The configuration of a horizontal antenna on ground or facing the ground with
respect to the normal configuration of a transmitting antenna. The simulation
software does not consider the reflection from the ground in the case where an antenna
is facing towards the ground. The thickness of a reader antenna is not considered.
The MATLAB software developed described in Section 3.7 was updated to handle hori-
zontal antennas with any elevation. One additional input parameter is required, which
is the elevation level with respect to a reference plane. The default reference plane is
the x-y plane 1 m above the ground shown in Fig. 3.7.
By replacing the rightmost antenna in Fig. 3.11 with a horizontal antenna (0 m eleva-
tion), the simulation result, shown in Fig. 3.13 when compared with Fig. 3.11, shows
that the boundary to deploy a third reader antenna operating in the same frequency
channel moves slightly closer to the first and second antennas, which improves the
situation slightly.
3.9.4 Antennas Operating in Different Channels
All the discussions on the Regulatory Reader Shutdown (RRS) problem presented in
previous sections are on multiple reader antennas operating in the same channel. In
this section, the effect of readers operating in neighbouring channels will be investi-
gated in the context of RRS.
Page 57
3.9 Results and Discussion (RRS)
−124
−126
−129
−131
−133 −136 −138 −140 −143
−121
−119
−117
Figure 3.13. Simulation results showing the effect of horizontal antenna configuration. Re-
sults from simulation (vertical and horizontal axes in units of 10 m, representing the y
and x axes shown in Fig. 3.7, received power in dBW). The bold line is the approximate
safe boundary of the deployment of an isotropic radiator in a same channel following
LBT. This figure shows that mixture of antennas with different orientation will improve
the dense reader environment as compared to Fig. 3.11.
For the case of a multiple-interrogator environment (Section 3.6.3), a simple and con-
servative way to give a rough estimation of inter-channel interference is available if all
the readers are compliant to [35]. Referring to the transmit mask shown in Fig. 3.5(a), it
can be seen that if a channel is being used for interrogation, the leakage to its immedi-
ate neighbouring channels must be below 20 dBch. Hence, if the simulation is focused
on a particular channel, the interference from the next neighbouring channels can be
emulated by reducing the signal strength by 20 dB.
However, it is trickier for the case of a dense-interrogator environment. Referring to
Fig. 3.5(b), the transmit mask depends on Tari, which is the reference time interval for
a data-0 in Interrogator-to-Tag signalling, and is between 6.25 µs and 25 µs [35]. In the
case study presented in this section, the focus will be on the European dense reader
environment, where the bandwidth of a channel is 200 kHz. Hence, for simplicity, Tari
is chosen to be 12.5 µs. By doing this, the -30 dBch level in dense-interrogator mask
Page 58
Chapter 3 Path Loss and Position Analysis
shown in Fig. 3.5(b) will coincide with the immediate neighbouring channels and the
-60 dBch level will coincide with the channels two channels away from the channel of
interest.
It must be noted that these transmit masks provided by [35] are international guide-
lines and are superseded by local regulations if the local regulations impose a stricter
transmit mask. This may be the case in a European country where ETSI EN 302 208 [48]
is used as the local regulation. However, at the time when this research was carried
out, [48] was undergoing a number of changes. Hence it was decided to use the trans-
mit mask for a dense-interrogator environment from [35] to demonstrate the ability of
the developed software to handle simulations involving antennas operating in differ-
ent channels.
An example to demonstrate the use of the developed simulation software to handle
the case where antennas are operating in different channels is shown in Fig. 3.14. The
antenna placement in this example is the same as the case shown in Fig. 3.13, where
there are a total of two transmitting antennas, located at (50,50) and (200,50). The
antenna located at (50,50) is of normal orientation with maximum gain facing in the y
axis, while the antenna located at (200,50) is of horizontal orientation with maximum
gain facing the z axis and has a 0 m elevation (refer to Fig. 3.12). Note that the difference
in height between the horizontal antenna and the normal configuration antenna has
been taken into account in the calculation. Also, in this scenario, these two antennas
are operating in the immediate neighbouring channels of the channel interest. Hence
to identify the safe zone to deploy a third transmitting antenna (which is an isotropic
antenna for simplicity), which is to operate in the channel of interest (not the channel(s)
occupied by antennas at (50,50) and (200,50)), the signals from antennas at (50,50) and
(200,50) are reduced by 30 dB, following the transmit mask for a dense-interrogator
environment shown in Fig. 3.5.
To avoid confusion in forming a guideline, the two antennas shown in Fig. 3.14 are re-
simulated individually and included as Fig. 3.15 to identify the safe distance between
a first transmitting antenna with gain pattern shown in Fig. 3.4 and a second antenna
(an isotropic radiator) to adhere the LBT provision. Fig. 3.15(a) shows the simulation
Page 59
3.9 Results and Discussion (RRS)
−126
−128
−130
−132
−134 −136 −138 −140 −142 −144 −146
−124
−122
−126
−124
Figure 3.14. Simulation results for some antennas operating in different channels and point-
ing in different directions. Results from simulation (vertical and horizontal axes in
m, representing the y and x axes shown in Fig. 3.7, received power in dBW). Transmit
mask for dense-interrogator environment is used with Tari chosen to be 12.5 µs. Note
that both antennas are operating in an immediate neighbouring channel of the channel
of interest.
results of an antenna with normal configuration while Fig. 3.15(a) shows the simulation
results of an antenna with horizontal configuration with 0 m elevation. Note that for
the case of an antenna with horizontal configuration, the simulation x-y plane is with
an elevation of 1 m (refer to Fig. 3.7 for an illustration of the position of the x-y plane).
Table 3.5 shows the safe distances between a first transmitting antenna with gain pat-
tern shown in Fig. 3.4 and a second antenna (an isotropic radiator) for various sep-
aration between the channel of interest which will be used by the newly introduced
antenna and the channels used by the first (existing) antenna. For a transmission oper-
ation in the channel of interest, the antenna must be placed 50 m away in all directions
from a horizontal antenna operating in the immediately neighbouring channel. For
an antenna in the default configuration (Refer Fig. 3.12) operating in the immediately
neighbouring channel, the distance is around 180 m from the front, 45 m from both
sides and 30 m from the back.
Page 60
Chapter 3 Path Loss and Position Analysis
−123
−126
−130
−134 −138 −142
−119
(a) First antenna in normal configuration
−126
−133
−138
−143
−148
−153
(b) First antenna in horizontal configuration
Figure 3.15. Determination of safe distance in accordance to LBT when the second antenna
is an isotropic radiator. Results from simulation (vertical and horizontal axes in m,
representing the y and x axes shown in Fig. 3.7, received power in dBW). Results for
both antenna in normal and horizontal configurations are tabulated in Table 3.5. Note
that both antennas in (a) and (b) are operating in an immediate neighbouring channel
of the channel of interest.
Page 61
3.9 Results and Discussion (RRS)
If the second antenna is not an isotropic radiator but is instead an antenna with similar
gain pattern to the first antenna, the safe distance will then depend on the orientation
and placement of both antennas. Nonetheless, the safe distance can be easily deduced
from the graphical presentation of the results such as those presented in Fig. 3.15.
For example, to locate a 6 dBi antenna facing the horizontal antenna at position (200,50)
in Fig. 3.15(b) and with channel difference of 1, the boundary must be as low as -
132 dBW (-126dBW-6dBi). From Fig. 3.15(b), the distance from the -132 dBW boundary
to antenna at (200,50) is approximately 75 m. Note that the safe distance is 50 m when
the second antenna is just an isotropic radiator and not a 6 dBi antenna.
Table 3.5 is sufficient to serve as a guideline on antenna placement in adherence to
the LBT provision in an RFID deployment zone. The only restriction is that all the
readers must be able to be pre-programmed to operate only in a pre-selected frequency
channel. However, upon a more careful inspection, it is not hard to realise that the safe
distances shown in Table 3.5 are unreasonably large. Hence the enforcement of LBT
will bring unreasonable restrictions to RFID deployment.
Table 3.5. Safe distance for different antenna configuration when the second antenna is an
isotropic radiator in accordance to LBT. Distances for channel difference 1 obtained
from Fig. 3.15(a) and Fig. 3.15(b). Results for other cases of channel difference were
obtained through separate simulations.
Channel DifferenceAntenna (Normal Configuration) Horizontal Antenna
Front (m) Side (m) Back (m) (m)
0 1400 350 210 320
1 180 45 30 50
2 130 25 15 35
3 95 20 10 30
Page 62
Chapter 3 Path Loss and Position Analysis
3.9.5 A More Hostile Environment
All the previous simulations are performed based on the path loss model as shown in
(3.7), using n = 2.5 for near distance and n = 3.5 for far distance. Hence, the simula-
tion results are approximate indications of various real life situations. A more hostile
environment will have a higher n and the safe distances in Table 3.5 will be reduced.
On the other hand, a less hostile environment will have increased distances. However,
it does not mean that RFID will perform better in a hostile environment. It is just that
in a dense reader and hostile environment, reader placement would be less restricted
under the “Listen Before Talk” provision as compared with that for a less hostile envi-
ronment.
3.10 Real Life Application
This section explores some real life applications where the coverage of an RFID in-
terrogation zone can be improved. Similarly to Section 3.9, the focus will be on the
Regulatory Reader Shutdown (RRS) problem, though the same idea can applied to
reduce the Reader Interference with Tag Replies (RITR) problem at the same time.
3.10.1 A Checkout Counter
As mentioned in Section 3.9.3, the term “horizontal antenna” is referred to the config-
uration as illustrated in Fig. 3.12 with certain elevation (default of 0 m). However, it
is usually not feasible to place a horizontal antenna directly on the ground, except in
some very specific applications. For example, in checkout points of retail outlets, it is
more sensible to place the antenna on the checkout counter level rather than ground
level (Fig. 3.16). The simulation results presented for the horizontal antenna are still
valid at far distance in all directions. There will be a slight difference in the results at
near distance. Alternatively, since the simulation software can handle a horizontal an-
tenna with any elevation, a new set of simulation results could provide a more accurate
prediction of the signal strength as received by an isotropic antenna at any location of
interest relative to checkout counter.
Page 63
3.10 Real Life Application
1 m
Ground
Horizontal Antenna(0 m Elevation, Upwards)
CheckoutItems
Horizontal Antenna(1 m Elevation, Upwards)
Figure 3.16. Repositioning of a horizontal antenna. Lifting a horizontal reader on the ground
to counter level.
3.10.2 Reader Synchronisation
One simple method to solve the RRS problem and adhere to the European regulation
(LBT provision) at the same time is to synchronise all the readers. All the readers will
be caused to start to “Listen” at the same time, and assuming no other short range
device is operating, will detect no signal at all as no reader is transmitting signals.
Following the European regulation, if no signal is detected in the “Listen” period, the
“Listen” period will end at the same time for all the readers [48], and the readers can
start interrogating tags, as shown in Fig. 3.17. Using reader synchronisation, readers
can avoid the RRS problem caused by LBT completely. A complete treatment of reader
synchronisation can be found in Chapter 5.
However it should be noted that reader synchronisation does not prevent other forms
of collision problem in a dense reader environment, such as the Reader Interference
with Tag Replies (RITR) problem. To minimise the RITR problem (Section 3.5), proper
separation and orientation of reader antennas are required. For simple cases, Table 3.3
can be used as a guideline for placing readers. For more complex cases, simulations
are required for each individual cases.
Page 64
Chapter 3 Path Loss and Position Analysis
Reader 1
Reader 2
"Listen" "Talk"
"Listen"
Reader N "Listen"
"Talk"
"Talk"
Time
Figure 3.17. Reader synchronisation. All the reader antennas are preprogrammed to “Listen” at
the same time before entering the “Talk” stage.
3.11 Conclusion
A detailed analysis of RFID indoor propagation models has been performed and dis-
cussed. This effort aimed to create a comprehensive guide to allow fast and successful
deployment of large-scale RFID systems and to maximise the potential and benefits of
that deployment. This path loss model can be applied to predict the signal strength
of an RFID interrogation signal at a certain distance away from a transmitting an-
tenna. The model has been implemented using a simulation script written in MAT-
LAB. Exemplary simulation results were presented, together with some insights and
discussions. Safe distances between antennas were suggested and these results can be
used in RFID large-scale deployment areas to avoid the problem of Reader Interfer-
ence with Tag Replies (RITR) and the problem of Regulatory Reader Shutdown (RRS)
in European countries with “Listen Before Talk” (LBT) provision. It has been shown
that strict regulations (such as LBT) will in fact limit large-scale deployment of RFID
systems. Proper antenna positioning has been proven to be effective in minimising
the problem of RITR. Reader synchronisation has been introduced to avoid the RRS
problem (a complete treatment of reader synchronisation can be found in Chapter 5).
More field testing results can be found in Chapter 4, where discussions on operational
considerations in simulation and deployment of RFID systems are presented.
Special Note: The research presented in this chapter was carried out in year 2005, with findings
showing the provision of “Listen Before Talk” (LBT) will impede large-scale deployment of
Page 65
3.11 Conclusion
RFID systems. LBT is officially removed from the latest version of ETSI 302 208 in year
2007 [59].
Page 66
Chapter 4
RFID OperationalConsiderations
LARGE-scale radio frequency identification (RFID) deployment
is needed for efficient item identification in supply chains. To re-
duce cost and save time, simulations are often carried out before
actual implementation, especially when RFID is used in regions in which
strict regulations and standards must be adhered to. However, due to the
unpredictable environmental effects on radio propagation, simple simula-
tion results can be misleading and questions have been raised over the va-
lidity of many wireless simulations. This research study reviews, from the
point of RFID antenna deployment, the sources of error in wireless simula-
tions reported in some publications. Also, the idea of second carrier sensing
is investigated to reduce the Reader Induced Tag Confusion (RITC) prob-
lem, and Specific Absorption Rate (SAR) of a dense reader deployment is
discussed to ensure human safety.
Page 67
4.1 Introduction
4.1 Introduction
This chapter is based extensively on the foundation work presented in Chapter 3.
There will be some overlapping of material between this chapter and Chapter 3. This
is to avoid extensive cross-chapter referencing, but an effort has been made to keep the
repetition to a minimal level.
In Chapter 3, a path loss model was proposed to simulate a dense reader environment.
A simulation software was developed to provide guidelines in antenna placements.
The main aims were to reduce two collision problems in a dense reader environment,
namely the Reader Interference with Tag Replies (RITR) problem and the Regulatory
Reader Shutdown (RRS) problem.
A good simulation of RFID deployment is essential as such simulation enables visu-
alisation of RFID deployment prior to actual implementation. However, a simulation
result depends on the complexity of the model used for the simulation. A simple simu-
lation can often lead to incorrect conclusions while a complex simulation may consume
too much computational time to be feasible. Several papers in the literature have in-
vestigated the validity and credibility of wireless network simulations [60, 61, 62]. Of
particular interest is [60], where several common simulation assumptions, which con-
tribute the most to simulation inaccuracy, are identified.
In this chapter, common simulation errors, such as use of an inappropriate simula-
tion model, neglect of antenna gain, and misinterpretation of simulation results, will
be discussed in the context of RFID, along with the exploration of the challenges in
RFID simulations. The simulation software used in this chapter is the same simulation
software presented in Section 3.7. Simulation results obtained through this developed
software are compared with measurement results, and several suggestions have been
offered to minimise the possible sources of error.
This chapter is structured in the following way. At first, RFID EMC related background
information is discussed in the next section, followed in Section 4.3 by the brief intro-
duction of the EPCglobal C1G2 RFID protocol and its proposed transmit mask in a
dense reader environment. As mentioned before, Section 4.2 and Section 4.3 contain
Page 68
Chapter 4 RFID Operational Considerations
some overlapping material with Chapter 3, but this is to enhance readability of this
chapter and reduce the frequency of cross-chapter referencing.
Section 4.4 presents a path loss experiment while Section 4.5 looks into several com-
mon sources of error in wireless simulation, and provides suggestions and ways of
minimising them.
Besides the discussion of sources of simulation error, this chapter also investigates two
further matters relevant to RFID operational considerations. Section 4.6 offers some
insight into the idea of second carrier sensing in reducing the Reader Induced Tag
Confusion (RITC) problem, its threshold limit selection and its dependency on an-
tenna positioning while Section 4.7 investigates the Specific Absorption Rate (SAR),
required to ensure human safety in a dense reader environment. Section 4.8 provides
the conclusions of this chapter.
Early parts of this chapter up to Section 4.5 have been published in [63]. Section 4.6 has
appeared in conference as [64].
4.2 RFID EMC Background
The regulatory status for using RFID in the UHF spectrum around the world can be
summarised into two categories: governed under Frequency Hopping Spread Spec-
trum (FHSS), and governed under Listen Before Talk (LBT). By the end of 2005, it is
expected that 50 countries, representing 83 % of the global GNI (Gross National In-
come) will have RFID regulations in the UHF band [65]. It is essential to understand
the differences between each regulation before an RFID simulation on deployment is
carried out, especially on the allocated bandwidth and the maximum allowable radi-
ated power.
Page 69
4.3 RFID Protocol
4.2.1 Frequency Hopping Spread Spectrum (FHSS)
An example under this category is the USA FCC Title 47 Part 15.247, with operation
within the band 902-928 MHz [47]. This bandwidth is then subdivided into 50 chan-
nels of 500 kHz. Each channel has a minimum transmission time of 0.4 s. Maximum
transmit power is 1 W with a maximum of antenna gain of 6 dBi, giving a maximum
Effective Isotropic Radiated Power (EIRP) of 4 W. It is adopted mainly in North and
South America.
4.2.2 Listen Before Talk (LBT)
An example under this category is the European ETSI 302 208 [48] that has been adopted
by some European countries. It allocates the frequency band of 865 - 868 MHz for RFID
deployment. This frequency band is then divided into 15 sub-bands or channels; each
spans a total of 200 kHz. When a reader is operating at the maximum total radiated
power, which is 2 W ERP (Effective Radiated Power) or equivalently to 3.2 W EIRP,
only 10 sub-bands are available, while the remaining 5 are utilised as guard bands in
which a low ERP is allowed. Three guard bands (out of the total five) are located at the
lower end of the designated bandwidth with a maximum ERP of 100 mW and the two
remaining guards bands are located at the higher end of the designated bandwidth
with a maximum ERP of 500 mW.
The detailed discussion of LBT can be found in Section 3.6.4. The transmit power and
the corresponding threshold values are extracted from the above-mentioned ETSI doc-
ument and integrated into Table 3.2. It is reproduced as Table 4.1 here for convenient
reference.
4.3 RFID Protocol
The operation of an RFID system is also standardised to encourage wide spread de-
ployment. EPCglobal has produced ”EPC Radio-frequency Identity Protocols Class-1
Generation-2 UHF RFID Protocol for Communication at 860 - 960 MHz” [35], in short
Page 70
Chapter 4 RFID Operational Considerations
Table 4.1. UHF RFID reader effective radiated power with corresponding threshold values
for LBT. Threshold value obtained from [48]. (From Table 3.2.)
ERP (W) ERP (dBW) Threshold (dBW)
Up to 0.1 Up to -10 ≤-113
0.1 to 0.5 -10 to -3 ≤-120
0.5 to 2.0 -3 to 3 ≤-126
EPC C1G2, as the standard operational protocol. This comprehensive protocol includes
the detailed air interface between readers and tags and also standard commands for tag
interrogation. The section in the protocol which is closely related to this research is on
the transmit mask for dense-interrogator environments as shown as Fig. 3.5. A more
relaxed transmit mask is used for low reader density surroundings [35] but is not pre-
sented here, as the focus of this research will be in a dense reader environment. For a
narrow bandwidth channel (European 200 kHz channel), it is suggested in this protocol
that odd-numbered channels should be used for tag backscatter while even-numbered
channels will be used for reader interrogation signals. For a wide bandwidth channel
(USA FCC 500 KHz channel), all available channels can be used for reader interroga-
tion, as tag backscatter signals will be located at the boundaries of these channels.
4.4 Path Loss Measurement Experiment
A simple path loss measurement experiment was described in outline in Section 3.3.
This section presents more details on the equipment setup, experiment locations and
data analysis.
The main objectives of this experiment are:
1. To demonstrate the multi-path effect on in-building path loss.
2. To provide experimental data for the investigation of in-building path loss model,
and to suggest the best model for a particular room.
Page 71
4.4 Path Loss Measurement Experiment
4.4.1 Experimental Setup:
The equipment used during the experiment is shown in Fig. 4.1. The antennas used
are approximately circularly polarized and have a gain of 8 dBi. In this experiment, we
are using a signal with a low transmit power generated by the HP ESG-3000A signal
generator, and hence the transmitted signal is in compliance with regulations.
Much of the theory concerning reflections which can be found in Sections 4.4.3 and
4.4.4 assumes linear polarisation. For direct rays between the two antennas the circular
polarisation produced by the test antennas will suffer the same loss as would linear
polarisation. However circular polarisation produced by these antennas would have
its sense reversed on reflection from metallic surfaces, and should not then be detected
by the receiving antenna.
Although at first sight this may seem to be a significant error for those paths involving
reflection, the polarisation produced by the antennas is imperfect even on bore sight,
and will be highly elliptical at the radiation angles involved in many reflections. It will
be particularly so for those paths for which the ray tracing of Section 4.4.4 is performed.
Moreover, most of the measurements in our path loss measurements involve direct
rays.
So we will regard the analyses and measurements in Sections 4.4.3 to 4.5.1 as capable
of providing useful data on which path loss models may be constructed.
4.4.2 Room Grid
Measurements were carried on a specified grid as shown in Fig. 4.2. The grid was
created in Room N203, Engineering North, the University of Adelaide, North Ter-
race Campus, South Australia. This room contains metallic cupboards, drawers, and
wooden tables and is chosen as an experimental site to represent a typical storage area.
The room is divided into a grid system, with markers positioned 1 m apart. Two an-
tennas of known gain are used as the transmitter and receiver. These two antennas
are directly facing each other (with each at exactly the same height) in order to obtain
Page 72
Chapter 4 RFID Operational Considerations
(a) Cushcraft 8 dBi S9028PC directional antenna.
(b) HP 8594E spectrum analyser
9 kHz - 2.9 GHz
(c) HP ESG-3000A signal generator
250 kHz - 3000 MHz
Figure 4.1. Equipment used in path loss prediction experiment.
for each an orientation with maximum directivity. Several measurements are taken at
various positions for a chosen separation and averaged.
4.4.3 Experimental Results
The raw data obtained from the experiment is shown in Table 4.2. The explanation
corresponding to each column is as follow:
• r: The distance between the transmitting and receiving antennas.
• Tx-Pt: The location of the transmitting antenna with reference to the room grid
as shown in Fig. 4.2.
Page 73
4.4 Path Loss Measurement Experiment
Figure 4.2. Experiment grid for signal strength measurement. For measurement more than 8 m,
the receiver antenna is moved out of the room through the door shown. The antennas
are facing each other with maximum gain attained for all measurements. Each grid is
1 m × 1 m.
• Rx-Pt: The location of the receiving antenna with reference to the room grid as
shown in Fig. 4.2.
• Tx-Pw: The transmitted power (in dBm) as delivered by HP ESG-3000 signal
generator (Fig. 4.1(c)). The power level is set, prior to the experiment, to -20 dBm
and is used for the entire measuring session.
• Rx-Pw: The received power (in dBm) as measured by HP 8594E spectrum anal-
yser (Fig. 4.1(b)).
• Gain: The total antenna gain (in dBi) of both the transmitting and receiving an-
tenna (Fig. 4.1(a)) used in the experiment.
• Loss: The total losses caused by coaxial cables used in connecting antenna (Fig. 4.1(a))
to spectrum analyser (Fig. 4.1(b)) and antenna (Fig. 4.1(a)) to signal generator
(Fig. 4.1(c)) . The loss value presented in Table 4.2 is an assumed value (in dB).
• PL: The path loss (PL), in dB, calculated using the measured power, taking con-
sideration of antenna gain and cable loss: PL = (Rx-Pw) - (Tx-Pw) + (Gain) -
(Loss), where Rx-Pw, Tx-Pw, Gain and Loss are defined above.
• FSPL: Theoretical free space path loss (FSPL), in dB, of the distance r, calculated
using
FSPL =(
λ
4πr
)2
(4.1)
Page 74
Chapter 4 RFID Operational Considerations
• Diff: The difference between the PL and FSPL, calculated using (PL) - (FSPL).
• Ave: The average of the Diff for all the same r. For example, a total of six mea-
surements were taken at different locations for the case of r = 1. The six Diff
measurement results were averaged to obtain an Ave value. This Ave value is the
same for all the cases with the same r. For the cases where r = 7 and 8, as there is
less than four measurement results, “N/A” (indicating not available) was input
in the Ave column instead.
Table 4.2. Average prediction of path loss based on experimental results. Explanation of
column titles can be found in Section 4.4.3, with Tx-Pw = -20dBm; Gain = 16dBi; and
Loss = 3dB.
r Tx-Pt Rx-Pt Rx-Pw PL FSPL Diff Ave
1 B8 C8 -39.5 32.5 31.67 0.83 0.95
1 A8 B8 -39.3 32.3 31.67 0.63 0.95
1 B9 B8 -40.0 33.0 31.67 1.33 0.95
1 C9 C8 -39.3 32.3 31.67 0.63 0.95
1 A6 A7 -39.6 32.6 31.67 0.93 0.95
1 D7 D6 -40.0 33.0 31.67 1.33 0.95
2 A8 C8 -45.5 38.5 37.69 0.81 0.39
2 B9 B7 -46.7 39.7 37.69 2.01 0.39
2 C9 C7 -46.0 39.0 37.69 1.31 0.39
2 A6 A8 -42.7 35.7 37.69 -1.99 0.39
2 D7 D5 -44.5 37.5 37.69 -0.19 0.39
3 A9 A6 -46.8 39.8 41.21 -1.41 -1.07
3 D7 A7 -48.8 41.8 41.21 0.59 -1.07
3 A6 A9 -45.5 38.5 41.21 -2.71 -1.07
3 D7 D4 -46.8 39.8 41.21 -1.41 -1.07
3 C7 C4 -47.8 40.8 41.21 -0.41 -1.07
4 D9 D5 -50.0 43.0 43.71 -0.71 0.44
Continued on next page
Page 75
4.4 Path Loss Measurement Experiment
Table 4.2 – Continued from previous page
r Tx-Pt Rx-Pt Rx-Pw PL FSPL Diff Ave
4 C9 C5 -51.6 44.6 43.71 0.89 0.44
4 C7 C3 -53.9 46.9 43.71 3.19 0.44
4 A9 A5 -49.1 42.1 43.71 -1.61 0.44
5 D9 D4 -48.8 42.8 45.65 -2.85 -1.28
5 B9 B4 -52.5 45.5 45.65 -0.15 -1.28
5 A9 A4 -49.3 42.3 45.65 -3.35 -1.28
5 C7 C2 -53.9 46.9 45.65 1.25 -1.28
6 D9 D3 -53.0 46.0 47.23 -1.23 -0.03
6 A9 A3 -53.3 46.3 47.23 -0.93 -0.03
6 C8 C2 -55.5 48.5 47.23 1.27 -0.03
6 B9 B3 -55.0 48.0 47.23 0.77 -0.03
7 A9 A2 -55.4 48.4 48.57 -0.17 N/A
7 C9 C2 -56.8 49.8 48.57 1.23 N/A
7 B9 B2 -59.0 52.0 48.57 3.43 N/A
8 A9 A1 -61.0 54.0 49.73 4.27 N/A
4.4.4 Discussion
1. It is suspected that PL at all near distances would be slightly lower than the FSPL.
This may be due to the multi-path effect where the propagating waves transmit-
ted by the antenna are reflected by the surrounding objects. At certain locations,
constructive interference will take place and will result in a detection of a stronger
received signal as compared with a received signal in a reflection free environ-
ment (such as an anechoic chamber). However, it has to be noted that destructive
interference may also occur.
To illustrate this effect, we can consider the addition of two waves, with different
phase. The power expression is as shown below:
Page 76
Chapter 4 RFID Operational Considerations
Preceived = (acosβ1z + bcosβ2z)2
= a2cos2β1z + b2cos2β2z + 2abcosβ1zcosβ2z
If a = 1, b can have among others the values of 1, j,−1,−j. The |Preceived| is then
4, 2, 0, 2, giving an average Preceived of 2.
A different approach is to compute the expected value of Preceived as shown below:
E [Preceived] = E[(acosβ1z + bcosβ2z)2
]= E
[a2cos2β1z + b2cos2β2z + 2abcosβ1zcosβ2z
]= E
[a2cos2β1z
]+ E
[b2cos2β2z
]+E [2abcosβ1zcosβ2z]
=a2
2+
b2
2+ 2ab
12
E [cos(β1z − β2z) + cos(β1z + β2z)]
=a2
2+
b2
2
The first approach predicts that although there appears to be nulls due to reflec-
tions, on average a gain is expected. The second approach predicts on average
there will be a gain due to reflection.
2. From the Diff column in Table 4.2, it can be seen that at some distance away the
PL computed using the measured signal strength is constantly lower than the
FSPL. For example, when r is 3 m.
Figure 4.3 shows the measuring configuration from a side view. For simplicity,
only one reflected wave is considered, which is the reflection from the ground.
The reflection from the ceiling is weaker than the reflection from the ground be-
cause the measurements were carried out at an elevation of 1 m as shown in
Fig. 4.3. The room where the measurement took place is approximately 3 m in
height and the reflection path from ceiling is much longer than the reflection
path from the ground. However, reflections from the surrounding objects, such
Page 77
4.4 Path Loss Measurement Experiment
as metal cupboards, walls or metal shelves, can have large impact on the receiv-
ing signal, depending on the material of the objects and the path of reflection.
Although these reflections from the surrounding objects are neglected in the dis-
cussion below, inconsistency in measurements result (as seen in Table 4.2) is most
likely contributed by them.
1 m
r
D
21
Ground
2D2
Figure 4.3. Measuring configuration from side view. r is the distance between the transmitting
and receiving antenna, D is the total length of the reflection path.
Another assumption is that the angles θ1 and θ2 in Fig. 4.3 are the same. This
will not be the case if the ground surface is not flat or uneven. Again, for sim-
plicity and to have a predictable reflection path length, a flat and even ground is
assumed and hence those angles are the same.
The results are shown in Table 4.3. It is very possible that the difference between
PL and FSPL is dependent on the difference in wavelength between the incident
path length and reflected path length. A difference of half a wavelength between
the incident path length and reflected path length, after subtraction of full wave-
lengths will most probably result in higher PL, while a difference of 0 wavelength
will result in lower PL.
For example, for r = 3 or r = 5, the difference in wavelength is between 0.1 and
0.2, and hence the computed PL values based on experiment measurements are
on average lower than the FSPL. On the other hand, for r = 2 or r = 4, the differ-
ence in wavelength is between 0.4 and 0.6 and the computed PL values based on
experiment measurements are on average higher then the FSPL.
Page 78
Chapter 4 RFID Operational Considerations
Table 4.3. Difference in wavelength between direct distance, r, and reflected path, D. An-
tennas are mounted following Fig. 4.3. The signal generator is transmitting a signal at
915 MHz with a wavelength of approximately 0.328 m. r is the distance between the
transmitting and receiving antennas in metre, D is the length of the reflection path in me-
tre, D-r is the difference between D and r in metre, and wl is D-r in terms of wavelength,
Ave is from Table 4.2 and is the average Diff values as explained in Section 4.4.3.
r D D-r wl Ave
1 2.236 1.236 3.768 0.95
2 2.828 0.828 2.527 0.39
3 3.606 0.606 1.848 -1.07
4 4.472 0.472 1.433 0.44
5 5.385 0.385 1.173 -1.28
6 6.325 0.325 0.991 -0.03
However, this argument is not strongly supported by measurements for r = 6.
From Table 4.3 and based on the pattern shown by r = 2, 3, 4 and 5, as explained
above, the PL values should be lower than FSPL. This inconsistency is most prob-
ably caused by reflection from nearby objects. For r greater than 6 m, too few
measurements are taken to form any conclusion. For a measurement more than
8 m, the transmitting and receiving antennas are separated by walls and the ef-
fect of walls must be taken into account. It is assumed that walls will decrease
the measured signal strength and hence increase the computed PL value.
This experiment has achieved objective 1 but has yet to produce a model, which
can best describe the observation of multi-path effects, mainly caused by reflec-
tion. More measurements were carried out in different locations (Appendix F) to
provide further understanding and insight in this study.
Page 79
4.5 Sources of Simulation Error
4.5 Sources of Simulation Error
A good simulation program should be able take into account the EMC regulations and
protocol standards discussed in Section 4.2 and Section 4.3. Also, the common sources
of simulation error of a wireless system as outlined in [60] and [61] must be minimised.
4.5.1 Path Loss Model
The term “path loss” is widely used in the literature and is carefully explained by the
authors in Section 3.2 and in Ref. [56]. Often the path loss model in a simulation is con-
sidered as a simple function of distance. This is only true for free space propagation,
and is used for satellite communication simulations [58]. A typical RFID deployment
zone is a warehouse filled with commercial products and logically a more complex
model is required. Hashemi [66] has categorised all the path loss models found in the
literature, apart from the simple free space path loss, into four groups.
These four categories have been studied carefully. It is found that a path loss model
with variable environmental factor, n, is suitable for the case of RFID as discussed in
Section 3.2, and this path loss model is reproduced here for convenient referencing:
PL(dB) = PL(d0) + 10 × n × log10
(dd0
)(4.2)
where d0 is an arbitrary reference distance, n is the environment factor, d is the separa-
tion distance between two antennas and PL(d0) is the free space path loss for a distance
d0. The n value takes into consideration, path obstruction, reflection, absorption and
other attenuation effects introduced by the presence of objects inside a building.
Also discussed previously in Section 3.2, (4.2) does not consider the fact that the n
value will increase as the distance increases. The modified version of (4.2) is presented
in Section 3.4 and is reproduced here as (4.3):
Page 80
Chapter 4 RFID Operational Considerations
0
20
40
60
80
100
120
1 10 100Distance (m)
Pa
th L
oss
(d
B)
n=2
n=3
n=4
Eq. (5.3)
Measured
1st Wall Encontered
More Walls
Encontered
Figure 4.4. Plot of path loss using different models together with measured results..
PL(dB) =
⎧⎪⎪⎪⎨⎪⎪⎪⎩
PL(d0) + 10 × n1 × log10dd0
0 < d ≤ 8 m
C + 10 × n2 × log10dd0
d ≥ 8 m
(4.3)
where at 915 MHz with n1 = 2.5 and n2 = 3.5, PL(d0) = 32 dB and C = 22.97 dB (see
Section 3.4 for details).
The comparison with various models and experimental results is plotted in Fig. 4.4
that shows a path loss model in free space, three models based on (4.2) with n = 2, 3 or
4, one model based on (4.3) and also the practical measurements from experiment. It
can be seen that (4.2), with n = 2 and PL(d0) set to the value for free space, is actually
the same as free space path loss. As the n value increases, the path loss increases for a
fixed distance.
The practical measurement was set up as shown in Fig. 4.2 in a room as described
in Section 4.4.2. The measurement shows several things. The first is that within the
room, the path loss is less than the free space path loss generally by a few dB. This
we attribute to reflections within the room, which would be expected to produce this
effect. Secondly, the practical measurement results follow an approximately free space
path loss pattern until the first wall is encountered along the measuring path. As the
Page 81
4.5 Sources of Simulation Error
distance increases, more major obstacles appear along the path of measurement, and
the practical measurement results have shown path losses higher than a free space path
loss. Equation (4.3) best mimics the measured results. Although the error margin can
be up to 10 dB, it can be minimised, if more appropriate values of n1 and n2 are used
in (4.3). However, the n value changes from location to location. On-site measurement
must be carried out to determine the best n for that certain area before the values of n
are used in simulations. Walls have been shown to have a great impact on the value of
n. Within a room, one fixed n value can be used, but n must be increased when a wall is
encountered as the distance increases. Equation (4.3) offers quick path loss estimation.
More complex surroundings with more obstacles in nearby vicinity requires a better
model as discussed in next section.
4.5.2 Reflection, Refraction, and Diffraction
A path loss model with environmentally determined n (such as (4.3)) has a high level
of inaccuracy if the environment of interest is full of obstacles, such as walls, metallic
cupboards, and narrow corridors. For example, in Fig. 4.5, the path loss in (a) and
(b) will not be the same, and will not be predicted by (4.3). This is caused by wave
propagation effects such as reflection, refraction and diffraction. Also, the appearance
of obstacles, such as a wall, must be accounted for the calculation of path loss.
The ray tracing method is well explained in the literature, such as [67, 68]. It is based on
geometrical optics and complemented by geometrical theory of diffraction to take into
consideration reflection, refraction and diffraction. On the other hand, [69, 70] modify
(4.2) empirically to cater for reflection, refraction and diffraction.
We combined both of the above ideas to produce a hybrid path loss model. The ar-
gument is that ray tracing normally requires many traces. If we limit ourselves to not
more than 5 rays, with each ray using an empirical model [69, 70], it is hoped that a
simple and accurate model can be obtained.
Using the settings presented in Fig. 4.5 (b), we constructed two additional rays (P1 and
P2) apart from the direct ray from transmitting antenna (Tx) to the receiver antenna
Page 82
Chapter 4 RFID Operational Considerations
(a)
Ceiling
(b)
Tx Rx Tx Rx
Ground Ground
P
P
1
2
MetallicBox
Figure 4.5. Comparison of the received signal strength with and without the presence of a
nearby metallic box. Signal received by receiver antenna (Rx) in (b) is stronger than
the signal received in (a).
(Rx). The distance of P1 and P2 is measured and with (4.3), the signals’ strength arrived
at Rx through direct path, P1 and P2 were added together to obtain the signal strength
as detected at Rx. P1 and P2 are traced above and below the direct path. Using the same
method, an additional two rays are constructed at the two sides of the direct path
of the transmitted signal to the receiver. The addition of these five signals’ strength
results in the received signal strength as detected by the receiver antenna as shown in
Table 4.4. Also, predicted signal strengths using this hybrid method and using (4.3),
are compared with mean value of measured signal strengths from experiment. It is
assumed that 50% of the power is reflected in the calculation whenever a reflection
occurs. The results show that this hybrid method reduces the error margin.
Though this hybrid method is still in its infancy, the main point here is that wave prop-
agation characteristic (especially reflection) will have great impact on signal transmis-
sion and must be taken account in simulation planning.
Page 83
4.5 Sources of Simulation Error
Table 4.4. Comparison of measured and predicted signal strength. D is the distance between
the transmitting antenna (with -20 dBm power delivered to the antenna) and receiving
antennas when both of them are facing each other; Pmea is the measured signal strength;
Plin and Phyb are the predicted signal strength using Eq. (4.3) and the hybrid method
respectively; ηlin is the difference between Pmea and Plin, and ηhyb is the difference
between Pmea and Phyb.
D (m) Pmea (dBm) Plin (dBm) ηlin (dB) Phyb (dBm) ηhyb (dB)
1 -39.62 -40.94 1.32 -40.33 0.71
2 -45.08 -48.47 3.39 -46.67 1.59
3 -47.14 -52.87 5.73 -50.21 3.07
4 -51.15 -55.99 4.84 -52.79 1.64
5 -51.38 -58.41 7.03 -54.86 3.48
6 -54.20 -60.39 6.19 -56.60 2.40
7 -57.07 -62.07 5.00 -58.10 1.03
8 -61.00 -63.52 2.52 -59.43 -1.57
35 -79.27 -86.31 7.04 -81.58 2.31
50 -82.27 -91.73 9.46 -86.98 4.71
As a side note, fluctuations of signal strength along the direct path (shortest path) be-
tween the transmission and the receiver antennas is observed. This is caused by con-
structive and destructive interference by reflected waves. Simulation models presented
in this research are not able to predict this and it is left as future research.
4.5.3 Radiation Pattern of Antenna
The power transfer between any two antennas is dependent on the angle of orientation
between them [60], unless they are isotropic radiators. A typical RFID antenna is a
6 dBi gain directional antenna and an example of the radiation pattern of a typical
Page 84
Chapter 4 RFID Operational Considerations
RFID antenna is shown in Fig. 4.6. This is the same radiation pattern presented in
Section 3.6.2, where it is used in the simulation software developed in Chapter 3. It is
reproduced here for easy referencing.
Despite the fact that in actual RFID deployment, it is very difficult to ensure all station-
ary RFID antennas are mounted in the fashion intended, with exact orientation, eleva-
tion and angles, the radiation pattern of any type of antenna to be deployed must be
acquired and fed into RFID simulation program as accurately as possible. Experience
shows that even two antennas of a same model manufactured by a same company will
have a slightly different antenna gain patterns. However the difference is small and
hence, the simulation program can assume the same gain pattern.
30
210
60
240
90
270
120
300
150
330
180 0
6 dBi 10
0
−10
−20
Figure 4.6. Polar plot of the antenna gain of a directional circularly polarised RFID antenna
with a gain of 6 dBi. (From Fig. 3.4.)
4.5.4 Simulation Result Interpretation and Analysis
A simulation program will use a model to compute the strength of signal at a certain
distance away from any transmitting antenna. There are many ways of presenting
a simulated result, either numerically or graphically. Fig. 4.7 shows an example of
simulated results in a graphical form, obtained using simulation software developed
in Section 3.7. In this example, two directional antennas with gain pattern shown in
Fig. 4.6 are located at position (50,50) and (400,50), both facing in the y direction and
Page 85
4.6 Second Carrier Sensing
Boundary
Figure 4.7. Boundary zone in simulation results. Results from simulation (vertical and horizontal
axes in units of 10 m, representing the y and x axes shown in Fig. 3.7, received power
in dBW).
with 1 W input power. Note that simulation results represent the received power by
isotropic antennas.
However, in real life applications, as shown in [60], the reception of a signal does not
exhibit a sharp cliff. For example, the line in the top of Fig. 4.7 shows signal strength
of -109 dBW. This line does not represent a clear threshold line between a region with
signal weaker than -109 dBW and region with signal stronger than -109 dBW. This is
due to uncontrollable factors that present themselves in a real life scenario, such as
fluctuation of path loss and antenna gain. However, a “boundary zone” (Fig. 4.7),
rather than a boundary line, can be specified in the simulation result to give good
estimate of the overall system performance before an actual deployment is carried out.
The size of the boundary zone should reflect the previously measured uncertainties in
the propagation loss which have already stated to be a few dB.
4.6 Second Carrier Sensing
To maintain a low production cost, an RFID tag normally does not have the ability
to filter a valid interrogation signal in a channel from the other valid interrogation
Page 86
Chapter 4 RFID Operational Considerations
signals in adjacent channels. Hence the RFID tag can be “confused” when more than
one reader attempts to interrogate it. This problem is a type of collision problem and
is known as the Reader Induced Tag Confusion (RITC) problem (refer to Chapter 3 for
more discussion of collision problems). This situation is common in a dense reader
environment, such as a distribution warehouse.
Mitsugi first classified the “Listen Before Talk” (LBT) (refer Section 4.2.2) as first carrier
sensing and has suggested a second carrier sensing [71], to avoid the RFID tag’s being
“confused” by multiple interrogation signals, or in other words, to avoid the RITC
problem.
The flowchart of reader transmission which involves the use of second carrier sensing
in first carrier sensing enforced areas, is as illustrated in Fig. 4.8. Note that the first
carrier sensing part of this flowchart does not represent the whole picture of the first
carrier sensing (LBT) under the ETSI 302 208. For example, the case where a reader
antenna can choose to wait until the chosen channel is free for transmission after it
fails its first carrier sensing, is not shown in the flowchart. However, Fig. 4.8 does
represent a valid interpretation and a rational path of the LBT provision.
When a reader is ready to transmit, it chooses a channel and carries out LBT (or first
carrier sensing), to check whether there is any signal stronger than the regulated thresh-
old values in the channel of interest. If there is, it will choose a new channel and per-
form the LBT again. If there is no signal stronger than the threshold value, the reader
will carry out the second carrier sensing.
When the second carrier sensing is performed by a reader antenna, the reader antenna
is trying to use signals detected across the whole RFID band of operation as an approx-
imate estimate of whether the tags which this reader antenna is trying to interrogate,
will be “confused” by signals from other reader antennas or not. As mentioned before,
an RFID tag does not have the ability to differentiate signals of different frequencies.
Hence, to avoid this RITC problem, it is recommended that at a tag, the signal coming
from the intended reader antenna should be at least 15 dB higher than signals coming
from any other readers, to ensure a BER of at least 10−4.
Page 87
4.6 Second Carrier Sensing
After a reader antenna of interest passes its first and second carrier sensing, it can start
to use its chosen channel to interrogate tags. If this reader antenna fails its second
carrier sensing, it has restart the entire process, which is to choose a new channel and
start the first carrier sensing again. It is hoped that after reselecting a new channel and
re-performing the first carrier sensing and by the time the new second carrier sensing
is carried out, the sources of interference (mostly mobile or hand held reader antennas)
would have been moved away from the nearby area of the reader antenna of interest.
The main difference between the first carrier sensing and the second carrier sensing is
that the first carrier sensing only checks for signal strength in the channel of interest
while the second carrier sensing checks for signal strength for the entire RFID band.
The threshold value for second carrier sensing is noticed, in the results to follow, to be
always higher than the first carrier sensing.
Reader ready to transmit
First Carrier Sensing
Choose a new channel
Pass?YesNo
Second Carrier Sensing
Lower
than
Threshold?
Yes
No
Reader Transmits
Figure 4.8. Second carrier sensing with “Listen Before Talk” provision. Flowchart showing the
stages before and after second carrier sensing is carried out. After a reader has found
a free channel, using first carrier sensing, in which to send a signal, this reader will
scan the entire frequency band allocated for RFID systems under the local regulations.
If this reader detects any signal in the frequency band higher than a pre-programmed
threshold value, it will decide not to transmit until successful second carrier sensing is
achieved.
Page 88
Chapter 4 RFID Operational Considerations
Mitsugi has presented results of second carrier sensing on random antenna placement.
However, we believe that antenna positioning is not random in nature. In practice,
when antennas are mounted, they follow a certain common pattern, such as side-by-
side or corner mounting, as shown in Fig. 4.9.
d
d 1
d
d 1
A
A
B
B
A
B
dd 1
(a)
(b)
(c)
Figure 4.9. Common antenna positioning. (a) face-to-face (b) side-by-side (c) corner.
It is obvious that second carrier sensing is not required on a repeating basis in an en-
vironment where only fixed position antennas are deployed, as in this case, they will
either fail or pass the second carrier sensing most of the time. For example, in a room
with two operating RFID readers with one transmitting antenna each, no matter which
channel one reader antenna has chosen to interrogate RFID tags, the signal strength
level within this reader antenna caused by the other reader will be detected and com-
pared with the pre-defined threshold of second carrier sensing. Since those antennas
are fixed position antennas, the signal strength will either be below or above the thresh-
old of second carrier sensing most of the time. However, in an environment where
hand held or mobile readers are used frequently, second carrier sensing can prevent
RITC.
The aim of the research presented in this section is to devise a sensible way to deter-
mine appropriate values to be used as the thresholds of the second carrier sensing for
Page 89
4.6 Second Carrier Sensing
different tag positions and environments. The parameters of the reader positionings
for the investigation carried out are illustrated in Fig. 4.9.
Referring to Fig. 4.9, a tag is positioned at location X, at distance d directly in front of
antenna A. First, the signal from A (2 W ERP) received by the tag at X is calculated.
Then to have a BER of 10−4, a SNR (or CIR) of approximately 15 dB is needed. Hence,
the signal from B received by the tag at X must be at least 15 dB lower than the sig-
nal from A. From there, the minimum distance of antenna B from the tag for that d
is obtained. With this minimum distance, the minimum distance d1 between the two
reader antennas can be easily obtained. Using the path loss model shown in (4.2) with
a chosen environment factor n, the signal strength of the signal transmitted from an-
tenna B as detected by antenna A can be computed. This signal strength is used as the
threshold of second carrier sensing.
Hence when a tag is located distance d away from antenna A and antenna B is moved
closer than the just calculated d1, (Fig. 4.9 (a)), the second carrier sensing will prevent
antenna A from operation and will so avoid the RITC problem. If antenna B is at least
distance d1 away from antenna A, antenna B is allowed to transmit under the scheme
of second carrier sensing, as a BER of at least 10−4 has been ensured for the tag.
For cases (b) and (c) as shown in Fig. 4.9, the antenna gain pattern must be used to find
the optimum angle, θ, to have a suitable d1 and a suitable threshold value for second
carrier sensing.
4.6.1 Effect of the Position of Tags
It is highlighted in the previous section that the distance of the tag from its intended
reader antenna, d, is an important factor in deciding the suitable threshold value for
second carrier sensing. To observe the effect of the position of a tag, a simple path loss
model, (4.2), is used, instead of the suggested piece-wise linear path loss model for
RFID application, (4.3). The environment factor in (4.2), n, is set to 3.0 for all results
shown in this section.
Page 90
Chapter 4 RFID Operational Considerations
As the distance of the tag from antenna A increases, the minimum separation distance
d1 between antenna A and antenna B increases. Once d1 is obtained, the threshold
value corresponding to d1 is computed, the results being shown in Table 4.5. From the
results, it is concluded that the face-to-face orientation (Fig. 4.9 (a)) should be avoided,
as the minimum distance between the two antennas becomes relatively large, and this
creates a problem in a dense reader environment. A side-by-side orientation offers the
best results among these three orientations, with relatively low minimum separation
distance. This means more reader antennas can be positioned in an area to increase
RFID tag detection coverage.
Note that Table 4.5 does not imply whether a high threshold value is desirable or not.
The threshold values tabulated in that table simply correspond to the minimum sepa-
ration distances. In other words, to have a minimum d1, the chosen threshold values
for second carrier sensing cannot be lower than the values shown in the table. If not,
one reader antenna will be permanently shut off by the test of second carrier sensing.
4.6.2 Effect of Environment Factor
As discussed in Section 4.6.1, (4.2) is used for the study of second carrier sensing. To
observe the effect of environment factor, the tag distance from antenna A, d, is set to
3 m for all results shown in this section.
With the environment factor in (4.2), n, ranging from 2.5 to 4 with increments of 0.5, the
minimum d1 is calculated for each case together with the respective threshold value.
The results (shown in Table 4.6) are as expected. An incrementing n results in higher
path loss value, and hence a lower distance is required to attenuate the transmitted
signal to an acceptable level at the receiver side. Again, similarly to the analysis of the
effect of position of tags in Section 4.6.1, side-by-side orientation offers better flexibility
for antenna positioning, in that the interfering antenna may be positioned closer to the
antenna that is reading the tags.
Page 91
4.6 Second Carrier Sensing
Table 4.5. Threshold values for second carrier sensing with respect to tag read range for
different antenna orientations. d is the distance between a tag and the nearest reader
antenna; d1 is the minimum distance between two reader antennas to have BER of at
least 10−4 for the tag, while the respective threshold values for second carrier sensing
are shown in the last column of the table.
d (m) d1 (m) Threshold (dBm)
Face-to-face
1 4.16 -8.58
2 8.32 -17.61
3 12.49 -22.89
4 16.65 -26.64
Side-by-side
1 1.30 -33.42
2 2.61 -42.50
3 3.91 -47.77
4 5.21 -51.51
Corner
1 4.28 -25.40
2 8.56 -34.43
3 12.84 -39.61
4 17.12 -43.46
4.6.3 Combining First and Second Carrier Sensing
If second carrier sensing is to be applied in a real life situation together with the first
carrier sensing, there remains a question on how to apply the positioning guidelines
already presented in this thesis in a correct sequence. The guidelines for first carrier
sensing (i.e. LBT) are discussed and presented in Section 3.9 while the guidelines for
second carrier sensing are presented in Section 4.6.1.
Page 92
Chapter 4 RFID Operational Considerations
Table 4.6. Threshold values for second carrier sensing with respect to environment factor,
n, for different antenna orientations. d1 is the minimum distance between two reader
antennas to have BER of at least 10−4 for the tag, while the respective threshold values
for second carrier sensing are shown in the last column of the table.
n d1 (m) Threshold (dBm)
Face-to-face
2.5 14.94 -19.36
3.0 12.49 -22.89
3.5 11.05 -26.52
4.0 10.11 -30.20
Side-by-side
2.5 4.28 -45.79
3.0 3.91 -47.77
3.5 3.64 -49.64
4.0 3.45 -51.51
Corner
2.5 15.24 -36.63
3.0 12.84 -39.61
3.5 11.45 -43.06
4.0 10.55 -46.63
Before going into an in-depth discussion, to avoid confusion, a few important points
are worth highlighting:
1. The antenna positioning guidelines for second carrier sensing depend on the
required tag read range, but the antenna positioning guidelines for first carrier
sensing do not.
2. The threshold of second carrier sensing was observed in our experiments to be
always higher than first carrier sensing. If this had not been the case, the first
carrier sensing would have been redundant. As the second carrier sensing checks
Page 93
4.6 Second Carrier Sensing
for signal strength for the entire RFID band, which includes the channel chosen
for interrogation, if the threshold of second carrier sensing had been be lower
than first carrier sensing, it would have incorporated the function of first carrier
sensing with a stricter threshold.
3. As things currently stand, a failure in the test of first carrier sensing does not
imply a failure in the test of second carrier sensing, and vice versa.
Also in the beginning, two pieces of essential operating information are required, which
are the distance between a tag and the intended reader antenna, and the position and
orientation of reader antennas. The tag read range requirement depends on the ap-
plications of the RFID system, and may vary from 0.1 m up to 10 m. Some common
antenna orientations have been shown in Fig. 3.9.
Once the tag read range requirement, and the position and orientation of reader anten-
nas have been decided, there exist only two logical ways to proceed:
• The threshold value of the first carrier sensing is used as the positioning guideline
before that of the second carrier sensing.
Using the software developed in Section 3.7 with the technique in simulating an-
tennas operating in different channels in Section 3.9.4, antennas are ensured to be
able to operate (pick a free channel successfully) under first carrier sensing. Then,
the required tag read range is used to compute the minimum distance between
antennas, d1, to have at least BER of 10−4. If any two antennas are positioned
nearer than d1, they will be relocated at a further distance away (distance of at
least d1). The final step is to determine for each antenna the threshold value for
second carrier sensing that corresponds to d1, as these numbers will need to be
programmed into the readers connected to the antennas.
• The threshold value of the second carrier sensing is used as the positioning guide-
line before that of the first carrier sensing.
The tag read range is used to compute the minimum distance between antennas,
d1, to have at least BER of 10−4. If any two antennas are positioned nearer than
Page 94
Chapter 4 RFID Operational Considerations
d1, they will be relocated to a further distance apart. Then, using the software
developed in Section 3.7 with the technique in simulating antennas operating
in different channels in Section 3.9.4, antennas are ensured to be able to oper-
ate (pick a free channel successfully) under first carrier sensing. If any of the
reader antennas fails to obtain a free channel for transmission, antennas need to
be repositioned again. The final step is to determine the threshold values for
second carrier sensing that are required by each reader.
It can be seen that both ways are almost the same, with a difference in the sequence of
steps. Both ways determine the threshold value for second carrier sensing as the last
step, as this step is only carried out after the position of all the antennas involved has
been fixed and no longer requires any changes. In short, it is important to determine
the safe distance between reader antennas in the setting up of an RFID system with the
enforcement of first and second carrier sensing. After antennas have been placed in
the designed locations, the threshold values for second carrier sensing correspond to
the safe distance and intended tag read range are input into readers.
Once all the fixed position antennas are in place and pass the selected second carrier
sensing criteria, they should pass the second carrier sensing with these criteria unless
they are in a scenario where hand held or mobile reader antennas are present.
In the scenario where a hand held or mobile reader antenna enters an RFID enabled
environment with the enforcement of first and second carrier sensing, the readers will
readjust the channel distribution among antennas using first carrier sensing. However,
if ever a hand held or mobile reader antenna moves too close to an operating fixed
positioned antenna, that antenna will be switched off by second carrier sensing, to
avoid the RITC problem.
In a dense reader environment where antenna placement is very restrictive for both
first and second carrier sensing, fine-tuning methods for RFID systems are suggested
by [72], and are discussed in Section 5.6.
Page 95
4.7 Investigation of Specific Absorption Rate (SAR)
4.7 Investigation of Specific Absorption Rate (SAR)
Specific Absorption Rate (SAR) is a measure of the rate at which radio frequency (RF)
energy is absorbed by the human body when exposed to radio frequency. This section
investigates the human radio frequency exposure level caused by 64 RFID reader an-
tennas positioned in the systematic array shown in Fig. 4.10. The focus is on the public
exposure (strictest level), and not the occupational exposure.
4.7.1 SAR Level for UHF RFID Operation
The safety levels discussed in this section are obtained from “Radiation Protection
Standard: Maximum Exposure Levels to Radiofrequency Fields - 3 kHz to 300 GHz,
Radiation Protection Series Publication No.3”, published by Australian Radiation Pro-
tection and Nuclear Safety Agency (ARPANSA) in 2002 [73].
The Australian RFID band spans from 920 MHz to 926 MHz, with a total of 12 channels.
In the case of public exposure, the maximum allowable power density, S, (in W/m2)
for frequency between 920 - 926 MHz, can be computed using the formula:
S =f
200(4.4)
where f is frequency in MHz.
From (4.4), a higher frequency will result in a higher maximum allowable power den-
sity. To have a conservative estimation, we chose a lower frequency to have a stricter
maximum allowable power density, which is 900 MHz. The maximum allowable
power density is then 4.5 W/m2.
4.7.2 Experiment Site
The 64 reader antenna array was used in an experiment to examine the interference
of a dense RFID reader environment to Australian GSM mobile telephone channels
Page 96
Chapter 4 RFID Operational Considerations
in Australia. This investigation of the SAR is to ensure human safety throughout the
duration of this experiment. The experiment site is shown in Fig. 4.11.
The main purpose of this study is to ensure the SAR 50 m away from the 64 anten-
nas array falls below the regulated maximum allowable power density. As shown in
Fig. 4.11(b), there is an office building 50 m in front of the test site. Also, it is the aim
of the study is to find out what is the safe distance from the 64 reader antennas array
when the maximum allowable power density is 4.5 W/m2.
4.7.3 Simulation of Power Density
As discussed in Section 4.7.2, the 64 antenna array is facing an office building. This
building is the location of interest and is 50 m away from the antenna site. Hence, 64
reader antennas were positioned as shown in Fig. 4.10 in the simulation software and
the simulated power density, Ssimu, is obtained at positions 50 m in front, 50 m on each
side and 50 m behind the 64 antennas.
The simulation of power density for the investigation on SAR is different from the
simulation of received power by a receiver antenna as presented in Section 3.7. This
is due to the fact that a human body is different from a reader antenna. It will absorb
RF power of different frequencies and the effect is cumulative. For example, a person
at a location exposed to 1 W/m2 contributed by a 900 MHz signal and to 1 W/m2
contributed by a 915 MHz signal, is in fact exposed to 2 W/m2 of power density.
This insight is consistent with the statement from [73] regarding the issue of simulta-
neous exposure to multiple frequency fields, which states that in general, exposure to
frequencies above 10 MHz are considered to be additive.
However, the software developed in Section 3.7 can be used to perform the required
calculation, as explained in the next section. In addition to this approach, a study of a
worst case scenario is presented in Section 4.7.5.
Page 97
4.7 Investigation of Specific Absorption Rate (SAR)
50 m
50 m
50 m
Front Measuring Point
Back Measuring Point
SideMeasuringPoint
Figure 4.10. Arrangement of 64 reader antennas for SAR simulation. 64 antennas are arranged
in 4 lines, 16 antennas each. Adjacent lines are 2 m apart. Adjacent antennas in each
line are 2 m apart. All antennas are facing in the same direction (front).
4.7.4 Maximum Power for Every Channel
For the investigation of SAR, a simulation was carried out using the software presented
in Section 3.7. This method requires only a single run of simulation, with all the anten-
nas radiating at maximum power level at a chosen channel.
Page 98
Chapter 4 RFID Operational Considerations
(a) Side View (b) Rear View
Figure 4.11. The experiment site of the 64 reader antenna experiment. All the reader antennas
are facing the same direction. Note that antennas used in this experiment are from
many different manufacturers. Nonetheless, they are all directional patch antenna.
In actual fact, the 64 antennas are not operating in a same channel, but each antenna
will randomly pick one of the available 12 channels under Australian regulations.
However, as discussed before, the RF power absorption effect of a human being is
cumulative regardless of the frequency of RF signals. Hence, it does not matter an an-
tenna is operating in which channel. For simplicity in simulation, all the antennas can
safely be regarded to be radiating at a same channel.
Simulation results are shown in Table 4.7. The results indicate that at 50 m away in
four major directions, the power density is far below the allowable 4.5 W/m2.
Table 4.7. Simulation results on SAR investigation using maximum power for every channel.
Ssimu is the simulated power density in W/m2.
Position Ssimu
50 m in front 0.00585
50 m behind 1.00 × 10−5
50 m on each side 4.05 × 10−5
This simulation does not consider the power contained by side-band leakage as it is
believed that the contribution of side-band leakage is very small. The explanation
Page 99
4.7 Investigation of Specific Absorption Rate (SAR)
to justify this point is based on the transmit masks presented in Section 3.9.3. The
dense-interrogator environments transmit mask (Fig. 3.5(b)) dictates that a side-band
leakage to the next immediate channel must be 30 dB below the signal strength of
the carrier signal. Hence, even when there exists a side-band leakage, its contribution
to the total radiation from an antenna will be more than 1000 times lower than the
contribution from that antenna within the channel in which it is radiating. For an area
with lower RFID reader density, the multiple-interrogator transmit mask (Fig. 3.5(a))
is used instead of the dense-interrogator transmits mask. The side-band leakage must
then be lower than 20 dB, and is still 100 times lower than a RFID signal operating in
that channel of interest.
Hence, even though the method of using maximum power slightly under-estimates
the power density, it provides a quick estimate of the power density at any location of
interest.
4.7.5 Worst-Case Scenario
The simulation results from Section 4.7.4 showed that it is safe at 50 m away from the
test site, with the method of using maximum power providing a slightly under but
quick estimation on power density. Hence it is in the interest of public safety that a
study of a worst-case scenario was carried out. Though physically unlikely or even
impossible, the following assumptions were made:
1. All antennas are represented by a single node and hence the resultant EIRP is
256 W EIRP (64 × 4 W EIRP).
2. Minimum path loss is used, which is the free space path loss.
3. There is a ground plane below resulting in perfect reflection.
4. The reflected wave is as strong as the direct wave and hence the signal is doubled
and the power density per unit area is quadrupled at the positions of interest.
Page 100
Chapter 4 RFID Operational Considerations
Table 4.8. Calculation results of the worst-case scenario.
Distance (m)With No Reflection Maximum Reflection
Power Density (W/m2) Power Density (W/m2)
1 20.37 81.48
2 5.092 20.37
3 2.263 9.053
4 1.273 5.092
5 0.815 3.259
6 0.566 2.263
7 0.416 1.663
8 0.318 1.273
9 0.251 1.006
10 0.204 0.815
15 0.091 0.362
20 0.051 0.204
25 0.033 0.130
30 0.023 0.061
35 0.017 0.067
40 0.013 0.051
45 0.010 0.040
50 0.008 0.033
55 0.007 0.027
60 0.006 0.023
Table 4.8 and Fig. 4.12 show the results of the worst-case scenario using the above
stated assumptions 1,2 and 3, and sometimes assumption 4. In the worst case, the safe
distance is 4 m away, but the public is not expected to enter this region. The public will
be in neighbouring buildings.
The locations of interest are located 50 m away from the 64 reader antenna array. The
power density level in the worst-case scenario at these locations is approximately 1135
Page 101
4.7 Investigation of Specific Absorption Rate (SAR)
0.00
0.01
0.10
1.00
10.00
100.00
0 10 20 30 40 50 60
Distance (m)
Po
wer
Densit
y (W
/m2)
No Reflection
Max Reflection
4.5 W/m2
Figure 4.12. Plot of simulation results on SAR investigation. Plot of Table 1, highlighting
the safe distance in front of the 64 closely spaced reader antennas but with and
without reflection from a ground plane. In the worst case using the maximum reflection
assumption, the safe distance approximately 4 m.
of the allowable power density level in the frontal direction and much less in other
directions.
4.7.6 Recommendations on SAR
The recommendations on SAR issue of the 64 reader antenna deployment:
1. Simulation results show the power densities 50 m away in four major directions
are much lower than the maximum allowable power density.
2. The worst-case scenario shows that the safe distance in the most power intense
direction is 4 m. At 50 m away, The power density level in the worst-case scenario
at 50 m away is approximately 1135 of the allowable power density level.
3. The occupants below the roof would be shielded from the metallic roof or metallic
water proofing layer.
Hence it is safe to the public to carry out the 64 reader antennas experiment according
to the maximum exposure level as recommendation by ARPANSA.
Page 102
Chapter 4 RFID Operational Considerations
4.8 Conclusion
This research study has suggested ways of minimising the common errors that arise
from ignoring multi-path propagation, antenna orientation and pattern, and variation
of path loss as obstacles are traversed, in the simulation of RFID antenna deployments.
Some errors are unavoidable, but the extent estimated from empirical measurements
can be shown to be not severe. Hence it is essential not to interpret any simulation re-
sult without bearing all the possible errors and underlying assumptions in mind. This
study hopes to provide a clear view of the pitfalls for researchers interested in RFID
simulator development. Also, this study has highlighted EMC regulations which are
essential in understanding the actual implementation of an RFID system, in order to
produce a sensible simulator, which will definitely contribute to the vision of automat-
ing supply chains using RFID technology. Furthermore, the idea of second carrier sens-
ing was investigated to assist the deployment of RFID system. The SAR studies shows
that even with a large number of reader antennas located in a nearby area, and with
some penalising and unrealistic assumptions, the SAR exposure level is not hazardous
to the public at a relatively short distance away.
Page 103
Page 104
Chapter 5
Reader Synchronisation
FOR a dense RFID reader deployment, such as in a warehouse,
where hundreds of readers will be positioned in a building, the
interference between all these readers must be studied carefully
to avoid the disruption of logistics operations. Strict RFID regulations and
standards have been imposed, trying to address the problem of reader col-
lision and also the problem of RFID devices interfering with other devices
operating in the same and nearby frequency bands. However, these guide-
lines and regulations are not entirely friendly for dense RFID reader deploy-
ment; in some cases it is not possible to have a feasible RFID system while
adhering to these regulations. Hence, this chapter proposes the synchroni-
sation of RFID readers to enable successful dense RFID reader deployment.
A case study targeted at European operations is presented in this chapter
to illustrate the actual synchronisation of RFID readers in real applications.
Some fine-tuning methods are also suggested to further improve the per-
formance of readers in a high reader density population area.
Page 105
5.1 Introduction
5.1 Introduction
Radio Frequency Identification (RFID) has received much attention recently as it is
widely believed that RFID can revolutionise supply chain management, complement-
ing barcodes as the main object tracking system. Several major supply chain operators
and retailers, such as Wal-Mart in the USA, have deployed RFID systems in some of
their supply chains [36]. Initial test runs of RFID deployment show encouraging re-
sults [74], and hence large scale RFID deployment is planned. However, before any
successful deployment can be achieved, some RFID issues have to be resolved. One of
them is the RFID reader collision problem, which is the focus of this chapter.
The term “reader collision(s)” is discussed extensively in [75] and [76]. In this chapter,
reader collision is simply defined as the phenomenon where an interrogation signal
from a certain reader disrupts the communication between a tag and another reader,
and this reader collision problem is potentially magnified in a dense reader environ-
ment, such as in a warehouse. Various regulatory and standardisation bodies have
tried to regulate the operations of RFID readers. In this research, the ETSI 302 208 as
introduced by the European regulatory body and the EPC Class 1 Generation 2 as rec-
ommended by EPCglobal are used as the basis of RFID reader operations. However,
as will be discussed in more detail in the latter part of this chapter, the restrictions that
are put on the operation of RFID readers are very strict, making it quite impossible to
have an uncoordinated large scale deployment of RFID readers. Hence, this chapter
further developes the idea of RFID reader synchronisation, to enable good RFID per-
formance in a dense reader environment, while adhering to strict regulations. The next
section introduces the ETSI 302 208 and EPC Class 1 Generation 2 Protocols and their
impact on RFID reader deployment. Section 5.3 explains the concept of RFID reader
synchronisation and how it adheres to strict regulations. Section 5.4 suggests possible
ways of implementing an RFID synchronisation system. A case study on RFID reader
synchronisation is presented in Section 5.5. Ways of fine-tuning RFID reader position-
ing are discussed in Section 5.6. Variations of possible reader synchronisation schemes
are presented in Section 5.7. Sections 5.8 offers views on the current progress of reader
Page 106
Chapter 5 Reader Synchronisation
synchronisation standardisation in Europe, followed by conclusion in Section 5.9. This
chapter is based extensively on [77] and [78], and is published as a book chapter in [79].
5.2 Background
5.2.1 ETSI 302 208
In the European Regulation as outlined in ETSI EN 302 208-1 V1.1.1 (2004-09) with “Lis-
ten Before Talk” (LBT) provision [48], a reader must “listen” and confirm that a partic-
ular channel is not occupied before it can use that particular channel to interrogate any
tag. The detailed discussion of LBT can be found in Section 3.6.4. The transmit power
and the corresponding threshold values are extracted from the above-mentioned ETSI
document and integrated into Table 3.2. They are reproduced as Table 5.1 here for
convenient reference.
Table 5.1. UHF RFID reader radiated power with corresponding threshold values for LBT.
Threshold value obtained from [48]. (From Table 3.2.)
Channel ERP (dBW) Threshold (dBW)
Up to 0.1 Up to -10 ≤-113
0.1 to 0.5 -10 to -3 ≤-120
0.5 to 2.0 -3 to 3 ≤-126
5.2.2 EPC Class 1 Generation 2 Protocol
“EPC Radio-frequency Identification Protocols Class 1 Generation 2 UHF RFID Pro-
tocol for Communication at 860 MHz - 960 MHz” [35], in short EPC C1G2, is the
standard protocol developed by EPCglobal for RFID devices for use within the sup-
ply chain. This protocol outlines the air interfaces and commands between an RFID
reader and an RFID tag. It also includes the spectrum management for RFID operation.
Page 107
5.2 Background
Frequency hopping or frequency agile systems are the suggested techniques. An allo-
cated frequency band, as allowed by local regulatory body, is divided into sub-bands
or channels. A reader will only use a certain channel for communication, not the en-
tire allocated frequency band. EPC C1G2 covers both dense reader mode and multiple
reader mode; multiple reader mode is for an environment where the number of simul-
taneously active readers is modest relative to the number of available channels while
dense reader mode is for an environment where the number of simultaneously active
readers is comparable to or more than the number of available channels. This docu-
ment only focuses on dense reader mode. In dense reader mode, for narrow bandwidth
(European 200 kHz) channels, it is suggested in this protocol that odd-numbered chan-
nels should be used for tag backscatter while even-numbered channels will be used for
reader interrogation. For a wide bandwidth channel (USA FCC 500 kHz channel [47]),
all available channels can be used for reader interrogation, as tag backscatters will be
located at the boundaries of these channels.
5.2.3 Problem in Dense Reader Environment
With the implementation of ETSI 302 208 and EPC C1G2, it is clear that when a reader
is operating at a certain sub-band or channel, this reader will effectively prevent other
readers from using that channel within an unacceptably large area. Section 3.9.3 has
presented detailed discussions and analysis on this matter and Table 5.2 (reproduced
from Table 3.5 for easy reference) summarises the minimum distance (calculated using
a piece-wise path loss model with variable environmental factor) between two anten-
nas connected to readers before one antenna operating at a certain channel will pre-
vent the other antenna from using that channel. It should be noted that these results
are obtained using a 0 dB isotropic receiving antenna, and do not represent any real
life situation, as a typical RFID antenna will be a directional antenna. Nonetheless, the
data presented in Table 5.2 gives sufficient evidence that a low threshold value for the
LBT as specified in ETSI 302 208 is severe enough to impede the reader deployment in
a dense RFID reader system.
Page 108
Chapter 5 Reader Synchronisation
Table 5.2. Safe distance for different antenna configuration when the second antenna is an
isotropic radiator in accordance to LBT. (From Table 3.5.) Distances for channel
difference 1 obtained from Fig. 3.15(a) and Fig. 3.15(b). Results for other cases of
channel difference were obtained through separate simulations.
Channel DifferenceAntenna (Normal Configuration) Horizontal Antenna
Front (m) Side (m) Back (m) (m)
0 1400 350 210 320
1 180 45 30 50
2 130 25 15 35
3 95 20 10 30
5.3 Reader Synchronisation
Under the concept of reader synchronisation, all the RFID readers in a certain area,
for example all the readers in a warehouse, are networked together through a cen-
tral control unit. The connection method can be the common Ethernet connection, or
equivalent, and will be discussed in the next section. Since all the readers are linked
together, physically or wirelessly, they can be directed to execute commands at a same
time. Also, they can be assigned channels dynamically, so that the spectrum manage-
ment is optimised while the reader collision is minimised.
European regulation allows ten channels when maximum radiated power, 2 W ERP, is
used. Following the recommendation of EPC C1G2, under dense reader mode, five of
them, the even-numbered channels, are used for reader interrogation. All the readers
are “Listen Before Talk” compatible. They are configured to start to “Listen” at the
same time, and then at the end of the listen period, they can all synchronously start
to “Talk” as shown in Fig. 5.1. This is due to the fact that according to ETSI 302 208,
if there is no signal detected in the intended channel of interest, the “Listen” time is
fixed. Hence, all the readers, which start “Listening” at the same time, will start “Talk-
ing” at the same time. If a reader is turned on at a different time, or if a reader loses
Page 109
5.4 Actual Implementation
synchronisation that reader can be made to start again in synchronism with the rest of
the readers, after the last reader has finished its “Talk” session.
Reader 1
Reader 2
"Listen" "Talk"
"Listen"
Reader N "Listen"
"Talk"
"Talk"
Time
Figure 5.1. Synchronisation of all readers. All the readers start to “Listen” at a same time and
finish “Listen” at a same time too. (From Fig. 3.17.)
5.4 Actual Implementation
5.4.1 Connectivity
In actual implementation, the RFID readers must be able to communicate with each
other to enable synchronisation of the RFID readers. There are basically two ways to
connect all the readers; either using wired (physical) connection or using wireless con-
nection. A physically connected system or wired system cannot support mobile read-
ers. Also, a wired connection may suffer from data latency in the network. Ref. [80]
shows that time synchronisation in a wired network is possible, but will require ad-
ditional hardware and system reconfiguration. In the best case, the time difference
achievable can be better than 1 ms. In an actual implementation, synchronising sig-
nals may suffer latency in a busy network. However, since all the readers in a certain
vicinity can be linked as a local area network, the network congestion problem can be
reduced.
A wired system is often considered as a more reliable and a more secure communi-
cation method than a wireless communication. A wireless system signals through an
RF link. This link can use one of the five guard bands, mentioned in Section 5.2, for
sending a synchronising signal. A synchronising signal can be a signal with a special
Page 110
Chapter 5 Reader Synchronisation
pattern. A wireless system can also use any existing wireless protocol such as Blue-
tooth technology. Wireless synchronisation supports mobile readers but is inevitably
vulnerable to interference (signal integrity problem), and unauthorised signal sniffing
(security problem). Both connectivity methods have their own advantages and disad-
vantages. The decision in choosing either of these two methods is largely dependent
on the positioning of the LBT sensor, which is discussed below.
5.4.2 Positioning of LBT Sensors
Reader
LBT Sensor
Reader Reader...
Antenna
Figure 5.2. Centralised LBT system. Readers in a nearby surrounding are connected to one LBT
sensor.
Reader
LBT Sensor
Reader Reader...
Antenna
Figure 5.3. Localised LBT system. Each reader has its own LBT sensor.
An LBT sensor of an RFID reader is responsible for detecting signals in the channel
of interest prior to transmission in that channel. This LBT sensor must have a power
sensitivity level better than -126 dBW as specified in [48]. If not, this LBT sensor will
not be able to function efficiently in determining whether there exists a signal with
Page 111
5.4 Actual Implementation
a power level higher than the power level specified in regulations in the channel of
interest. An LBT sensor can use the RFID antenna used for transmitting and receiving
signals in the communication with RFID tags. A LBT sensor can also use a separate
antenna connected to an RFID reader. Also, several RFID readers could share an LBT
sensor within a close vicinity. This is also known as a centralised system. A localised
system is where each and every RFID reader has its own LBT sensor. A centralised
LBT system is as shown in Fig. 5.2. The LBT sensor will constantly monitor all the
channels allocated for RFID operation, and dynamically assign available channels to all
the readers connected to it. The central control system has to be configured during the
initial setup of the system. A fine-tuned centralised LBT system offers high reliability.
However, it requires additional network configuration, to connect all the readers to the
LBT sensor using either wired or wireless connection. Also, a centralised LBT system
will not be able to be implemented effectively when mobile readers are dominant in
the surroundings. This is because although the readers can communicate with the
centralised LBT sensor through a wireless link, it is very difficult for the centralised
LBT sensor to estimate the position of mobile readers, and hence is not possible to
allocate the best channels for mobile readers.
For example, if two mobile readers operate simultaneously in an enclosed area, there
is a probability that the two readers move near to each other at some time. The cen-
tralised LBT sensor may at that time allocate very nearby channels to those two read-
ers and serious interference between those two readers may occur. Also, if two nearby
areas are running on different RFID wireless networks and they are un-coordinated,
interference with each other can occur, and in the worst case, cause a complete system
shut down. The coordination of wireless networks in different premises will be time
and cost consuming.
A localised LBT system is as shown in Fig. 5.3. Each reader has its own LBT sensor.
The LBT sensor can either use a separate antenna (Fig. 5.3), or use the same antenna
a reader uses to establish communication with an RFID tag within its interrogation
zone. As compared to centralised LBT system, a localised LBT system with wireless
connectivity enables new readers to be easily integrated into an existing system, with
Page 112
Chapter 5 Reader Synchronisation
no additional cabling or setup needed. However, a localised LBT system has the prob-
lem of management of channel sharing, signal interference and possibly creation of
unwanted shielding. In actual fact, the connectivity of readers and the positioning of
an LBT sensor are closely related. In [81], a wired system and a centralised LBT are
linked together as one configuration, while a wireless system and a localised LBT are
linked together as another configuration.
5.4.3 Antenna Positioning
Truck Docks
h
AntennaA
AntennaBBest
ReadZone
Moderate Read Zone
Figure 5.4. Antenna positioning in dock door. A typical antenna setting at dock door, with h
being the height of the antenna from the base of a dock door.
The positioning of RFID interrogation antennas depends primarily on the application.
Detailed operational considerations for the deployment of an RFID system are pre-
sented in [63]. In this chapter, only one example will be given, which is the dock door
situation, as it will be used in the case study in the next section. A dock door is usually
2 to 3 m in width and approximately 3 m in height. The most effective way to create
an RFID interrogation zone is to position two antennas at the sides of the dock door,
face-to-face and with an height elevation, h, as shown in Fig. 5.4. The height elevation,
h, mainly depends on the average height of objects being shipped through the dock
Page 113
5.5 Case Study
door. A normal choice of h is between 0.5 to 1 m. Also, antenna A and antenna B
normally will be using different channels for tag interrogation. However, if antenna A
and antenna B are operating at the same time, and a tag is located in the middle of the
dock door, the tag may be “confused” by the interrogation signals from both of the an-
tennas with the result that the tag is misread. This effect is known as the tag confusion
problem. The discussion of this issue is outside the scope of this chapter but a simple
solution to this is to alternate the operation of antenna A and antenna B every query
cycle.
5.5 Case Study
A case study on dense RFID reader deployment at the dock doors of a warehouse is
presented here. As shown in Fig. 5.5, the green coloured rectangles represent trucks
loading or unloading goods at the dock doors of a warehouse. Each door is around 3
m in width, and has two RFID antennas facing each other for tag interrogation.
Since all the readers are synchronised in a way described in Section 5.3, they will
start “Listening” at a same time and will be assigned a channel for interrogation at
the end of “Listen” period. The assignment of channels will be geographically influ-
enced. Two readers assigned to be operating in the same channel will be as far apart
as possible. Also, the neighbouring antennas will be using channels as far apart as
possible. As illustrated in Fig. 5.6, the spectrum is split into ten channels, all five of
the odd-numbered channels are reserved for tag backscattering while all five of the
even-numbered channels are assigned for reader interrogation. Fig. 5.5 shows how the
channel assignment is done. The antenna on the furthest left is using channel 2 for
interrogation. The next antenna on its immediate right is using channel 8, which is six
channels away. Channel 10, though it is the furthest channel away, is not chosen. This
is because the arrangement of 2, 8, 4, 10, 6 gives best channel separation between every
channel.
Page 114
Chapter 5 Reader Synchronisation
2 8 4 10 6 2 8 4 10 6 2 8Channel
Time"Listening"
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Figure 5.5. Alternating of “Listening” and “Talking” mode. All the reader antennas at the
dock doors “Listen” at the same time and “Talk” together at some other time.
Channel
frequency
1 2 3 4 5 6 7 8 9 10
Figure 5.6. Channelling of the allocated frequency spectrum. Odd-numbered channels will be
used for tag backscatter signals while even-numbered channel will be used for reader
interrogation signals.
5.6 Synchronisation Fine-tuning
Fine-tuning of a synchronised RFID system, as presented in this section, can be carried
out to further reduce the tendency of reader collision. The fine-tuning methods dis-
cussed below include the reduction of output power, the reduction of overall reader
talking time, the use of external sensors, the use of RF opaque or absorbing materials,
and the frequent rearrangement of channels allocations.
Page 115
5.6 Synchronisation Fine-tuning
Figure 5.7. Reduction of output power to fine-tune reader synchronisation. Estimation of
required radiated power given that maximum read range corresponds to maximum ra-
diated power.
5.6.1 Reduction of Output Power
Although up to 2 W ERP can be used in a single or small population reader environ-
ment, in dense reader populations this higher power may not be necessary. Currently,
a state of the art reader can read up to 10 m. However, normal reading operations
do not require such a read range. In the case study presented in Section 5.5, the dock
doors of the warehouse are around 3 m in width. Since two antennas are positioned
facing each other in every dock door, the read range required is also around 1.5 to 2 m.
By reducing the radiated power of readers, the minimum distance between two anten-
nas using the same channel can also be reduced, this reduction being beneficial in a
dense reader environment. Fig. 5.7 gives an approximation to the reduction of output
power. In the far field region, using the Friis equation, the power received is the in-
verse function of the square of distance (r2). If the maximum read range corresponding
to maximum radiated power (2 W ERP) of a RFID reader is known, we can compute
the required radiated power for a shorter read range. For example, if the maximum
read range of a reader is 5 m using 2 W ERP (shown in Fig. 5.7), and if a read range of
only 2 m is required, the required radiated power can be lowered to 0.32 W ERP. This
estimation may not be accurate in real life due to complex electromagnetic propagation
phenomena, such as reflection caused by the surroundings objects, but it demonstrates
that power reduction is a viable option.
Page 116
Chapter 5 Reader Synchronisation
5.6.2 Reduction of Overall Reader Talking Time
While it is possible to talk for 4 s, reader applications should be configured to talk
for only the time necessary to capture tag data. There is no optimum talking time. It
depends on the application and also the surroundings of the deployment zone. On-site
fine-tuning and measurements are needed before the reduction of talking time can be
carried out.
5.6.3 Use of External Sensors
Sensors can be used to turn RFID readers on only when tags are approaching thus fur-
ther reducing reader interference in that area. This will free up the channels allocated
for those antennas, and also to avoid unnecessary interference to other surrounding
reader antennas. For example, external sensors can be attached to the dock door in
the case study in Section 5.5. When the dock door is not in use, the designated RFID
readers would be switched off, as shown in Fig. 5.8. Optionally, the central control unit
can (as shown) dynamically shift the channels assigned for the antennas at door 3 to
door 4.
Channel
2 8 4 10 6 2 8 4 10 6
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Figure 5.8. Using sensors in an RFID system. Both the antennas at dock door 3 are switched
off when the absence of truck 3 is detected. This will free up the channels allocated for
those antennas, and also to avoid unnecessary interference to other surrounding reader
antennas.
Page 117
5.6 Synchronisation Fine-tuning
5.6.4 RF Opaque or RF Absorbing Materials
Another effective, but more expensive, way to reduce reader interference and collision,
is to utilise RF opaque or RF absorbing materials to contain the interrogating signal
within the designated zone of interrogation. For the case study presented in Section 5.5,
the use of such materials is shown in Fig. 5.9. Although there will still be some signal
leakage through the door openings, it would not have caused much interference. This
is due to the fact that the signal strengths at the sides of the antenna are relatively weak
as compared with the front of the antenna. According to [57], the gain at the side of
a typical RFID antenna is approximately 20 dB less than the gain at the front of the
antenna.
RF Absorbing Materials
2 8 4 10 6 2 8 4 10 6 2 8
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Tru
ck
Figure 5.9. Use of RF absorbing materials. All the antennas are separated using RF absorbing
materials. The antennas facing each other at the same door, are separated by at least
4 channels.
5.6.5 Frequent Rearrangement of Channels
Interrogating channels can be switched around every cycle of “Listen Before Talk”.
This is to prevent the jamming of the interrogation signal by any external noise. Fig. 5.10
shows a simple example on how the switching is done. There may be other more com-
plex switching methods involving higher artificial intelligence in the central control
unit, depending on the noise received from the surrounding environment, but these
await further development.
Page 118
Chapter 5 Reader Synchronisation
Channel
8 2 6 10 4 8 2 6 10 4 8 2Channel
2 8 4 10 6 2 8 4 10 6 2 8
Figure 5.10. Channel switching within antennas. The channels arrangement are flipped from
left to right. Note that any face-to-face antennas are operating at least four channels
apart.
5.7 Variation of Synchronisation
In the previous sections, suggestions on the implementation of a real life RFID reader
synchronisation system are presented, together with some deployment options, such
as the connectivity of all the readers. Also, fine tuning methods are presented. In this
section, some of the interesting variations of RFID reader synchronisation schemes
are presented. These variations may not be readily incorporated into the suggested
methods mentioned in previous sections, but are presented here for future reference
and for completeness.
5.7.1 Separation of Transmitting and Receiving Channels
For the RFID, full power operation (2 W ERP) as governed by ETSI 302 208, only ten
channels are available, as shown in Fig. 5.6. However, as discussed in Section 5.2, there
are actually fifteen channels available for RFID in total. Five of the fifteen channels,
though used as guard bands, can be used for RFID operation with reduced maximum
allowable radiated power. There are three channels located lower in frequency than
the normal ten channels. These three channels can only be operated below 100 mW
ERP. There are two channels higher in frequency than the normal ten channels. These
Page 119
5.7 Variation of Synchronisation
1 2 3 4 5 6 7 8 9 10 1112131415
1 2 3 4 5 6 7 8 9 10
Complete
2 W
100 mW 500 mW
Figure 5.11. The complete frequency band allocated for RFID operation in Europe. Only
ten channels available for maximum allowable power operation (2 W ERP).
1 2 3 4 5 6 7 8 9 10 1112131415
1 2 3 4 5 6 7 8 9 10TxTx TxTx
Rx Rx Rx Rx Rx
Figure 5.12. Separation of transmitting and receiving channels in the frequency band al-
located for RFID operation in Europe. All the Transmitting (Tx) channels are
separated by two Receiving (Rx) channels as compared to the complete frequency
band shown in Fig. 5.11.
two channels can be operated below 500 mW ERP. The complete frequency range for
RFID operation, with respective regulated power level is as shown in Fig. 5.11.
The channel numbering system shown in Fig. 5.6 is included in Fig. 5.11, along with
a new channel numbering system to simplify the discussion hereon. Channel 4, 7, 10
and 13 are assigned to be the reader transmitting channels while the tag reply channels
are the four channels beside the transmitting channels [82]. For example, transmitting
channel 4 uses channel 2, 3, 5, and 6 for tag reply. Although the transmitting channels
are reduced from a total of five down to four, the transmitting channels are placed
two channels away rather than one channel away from the neighbouring channels.
(Fig. 5.12). From the transmit mask shown in Fig. 3.5, an improvement of 5 dB can
be obtained. Hence with the reduction of interference between transmitting channels,
readers can be placed nearer to each other.
Page 120
Chapter 5 Reader Synchronisation
5.7.2 Separation of RFID and Non-RFID Signals
Another variation of synchronisation is to differentiate an RFID signal from a non-
RFID signal. A method using signal recognition is presented in [83]. The idea is that
all the RFID readers in a certain region could possibly be treated as a single entity
in the regulations as outlined in the ETSI 302 208. Hence it is only required in this
interpretation that signal interference between all the RFID readers and the rest of the
short-range devices be avoided. If this concept is valid, the interrogation signals of
RFID readers are not treated as a signal in a channel when an LBT test is carried out.
The main advantage of this method is that a lot of readers can be deployed in a small
confinement area. However, reader antenna positioning can become more challenging,
as all the readers can choose any channel for transmission as long as there is no other
type of short-range device around.
5.8 Updated Progress on Development of RFID Reader
Synchronisation
The regulation of RFID reader synchronisation in the European countries is governed
by ETSI. TG34: Radio Frequency Identification Devices (RFID), a technical work group
within ETSI has been field-testing late in 2006 in an operational distribution center us-
ing a synchronised RFID system [84]. The testing involved up to 36 adjacent portals
operating simultaneously. Using identical pallets comprising 63 “RFID unfriendly”
cartons, a read rate of better than 98.5 % is recorded. In a near future, RFID reader
synchronisation would be included in ETSI TR 102 436 “Electromagnetic compatibil-
ity and Radio spectrum Matters (ERM); Improved spectrum efficiency for RFID in the
UHF Band”, and be harmonised with the current EN 302 208. Also, with the standard-
isation of RFID reader synchronisation, the mandatory use of “Listen Before Talk”,
which is a deterrent to large-scale deployment of RFID system, will be lifted, allowing
a better-performance and better coverage RFID systems in Europe.
Page 121
5.9 Conclusion
5.9 Conclusion
This chapter has identified synchronisation of RFID readers as a mechanism to assist in
RFID reader deployment in dense reader environments. Some implementation meth-
ods, and several fine-tuning methods are also presented in optimising the performance
of a synchronised RFID system. As compared with conventional unsynchronised RFID
systems, a synchronised RFID system can offer more coverage, less reader collision or
interference, while strictly following the European regulations and the EPC C1G2 rec-
ommendation, and can, with variation of the normal operating procedure, deal also
with the effects of tag confusion. However, these benefits require the use of more
complex hardware and hence can marginally increase deployment costs. Reader syn-
chronisation has not been tested in a real situation, and hence will require future study
in this area.
Page 122
Chapter 6
RFID Tag Antenna Design
THIS chapter focuses on the design of RFID tag antennas, begin-
ning with the introduction of relevant antenna theory. The chal-
lenges in designing small RFID tag antennas are discussed, fol-
lowed by the presentation of the methodologies and tools used in the pro-
cess of simulation and fabrication of antenna prototypes. The design and
analysis of various types of HF and UHF RFID tag antennas, mostly indus-
trial driven, are presented at the end of this chapter.
Page 123
6.1 Introduction
6.1 Introduction
The design of an antenna for an RFID tag depends heavily on the application. In the
current RFID worldwide market there is a wide variety of RFID tags with very different
antennas of different sizes, shapes, materials, and operating frequencies. This is to
cater for the different needs of various applications.
Often, an RFID tag antenna designer will be given the specifications of which fre-
quency band the tag is to be operate in, the location on the product where the tag
is going to be attached, and the read range performance requirement for the tag.
There are standard design procedures for an RFID tag antenna. They are very useful
as a reference when the RFID tag antenna is to be attached to an object that does not
have a space constraint and which is to be deployed in an RFID friendly environment.
However, there are cases where the design of an RFID tag antenna is challenging.
Throughout the author’s research, some very challenging RFID tag antenna design
problems have been encountered, mostly presented by industry partners.
This chapter will contain discussion in Section 6.2 of the antenna theory involved in
designing an RFID tag, definition in Section 6.3 of the antenna parameters used in any
discussion throughout this thesis, and presentation in Section 6.4 of equivalent circuits
of an RFID antenna. With a firm grasp of the fundamental electromagnetic properties
of antennas, the discussion is extended in Section 6.5 to the limitations and challenges
of designing an RFID tag antenna.
The preparation of HF and UHF RFID tag chips used in the experiments is discussed
in Section 6.6 while the range of HF and UHF RFID readers and RFID reader antennas
are documented in Section 6.7.
Section 6.8 explains the fabrication process of designed tags presented in this thesis.
Also, the simulation and prototyping software is included as Section 6.9. The fabri-
cation process and testing methods of a prototype antenna are also discussed in Sec-
tion 6.9. A further investigation of antenna performance measurements can be found
in Chapter 8.
Page 124
Chapter 6 RFID Tag Antenna Design
Section 6.10 begins with the design of simple HF RFID tag antennas, and then moves to
a consideration of more complex HF RFID tag antennas which cater for specific needs
and specifications.
Section 6.11 is very similar to Section 6.10, except the focus is on the UHF RFID tag
antennas rather than the HF RFID tag antennas.
6.2 Antenna Theory
A major focus of this thesis is on the design of HF and UHF RFID tag antennas. Inten-
sive background studies based on [85] and [86] were carried out to establish a strong
foundation in antenna theory and to distinguish the difference between the design of
an HF RFID tag antenna and the design of a UHF RFID tag antenna. This section lists
several important points of the related antenna theory.
• Common terms in electromagnetism
The common terms in the study of electromagnetism are defined by The Interna-
tional System of Units [3] and are shown in Table 6.1.
Table 6.1. SI units for common terms used in antenna theory.
Electric field strength E Vm−1
Electric flux density D Cm−2
Magnetic field strength H Am−1
Magnetic flux density B Wbm−2
Permittivity ε Fm−1
Permeability µ Hm−1
• Maxwell’s Equations
The fundamental antenna theory is based on Maxwell’s equations, first intro-
duced by James Clerk Maxwell in unifying the theories of electromagnetism. The
Page 125
6.2 Antenna Theory
four Maxwell’s equations are the combination of Faraday’s Law, Ampere’s Law
as modified by Maxwell, Gauss’s Law for electric flux and Gauss’s Law for mag-
netic flux.
In differential form:
∇× E = −∂B∂t
(6.1)
∇× H = J +∂D∂t
(6.2)
∇ · D = ρ (6.3)
∇ · B = 0 (6.4)
• Sources and vortices
To visualise the electric, magnetic or electromagnetic fields, the concepts of a
source and a vortex are introduced. A source like field is as shown in Fig. 6.1(a)
and vortex like field is as shown in Fig. 6.1(b). Normally, an electric field is con-
sidered as a source like field while a magnetic field is considered as a vortex like
field.
(a) Source (b) Vortex
Figure 6.1. The difference between source and vortex.
• Boundary conditions
A brief treatment of electromagnetic boundary conditions is presented here, with
its important impact on RFID tag antenna design highlighted. A complete treat-
ment of boundary conditions can be found in [87].
Electromagnetic waves will undergo changes when they traverse from one medium
into another medium. If one of the media is a conductor, such as a metallic sur-
face, there can only exist a tangential magnetic fields and a normal electric fields
Page 126
Chapter 6 RFID Tag Antenna Design
at the boundary. Fig. 6.2 shows the changes on electric and magnetic field near a
metallic surface.
Charge
Metallic Surface
Wire
Figure 6.2. Boundary conditions affecting electric and time varying magnetic field. If the
metallic surface is a perfect conductor, there are only tangential magnetic fields and
normal electric fields at the boundary.
• Spherical Coordinate System
In the study of antenna theory, a spherical coordinate system is often used over
the conventional cartesian coordinate system. This is due to the nature of the field
pattern when it can be described in the simplest manner using a spherical coordi-
nate system. Fig. 6.3 shows the configuration of a standard spherical coordinate
system.
P( , , )r è
z
y
x
è
r
Figure 6.3. Spherical coordinate system. Any point in space can be specified by using (r,θ,φ)
coordinates shown above.
• Near field and far field
The analysis in this section is from [85] and [88] . The analysis applies to the case
where r > 0.
Page 127
6.2 Antenna Theory
The E and the H fields for an infinitesimal electric dipole can be expressed in
spherical coordinates as:
Hr = Hθ = 0 (6.5)
Hφ =β2 I0l4π
[j
(βr)+
1(βr)2
]e−jβrsinθ (6.6)
Er = ηβ2 I0l4π
[2
(βr)2 − 2j(βr)3
]e−jβrcosθ (6.7)
Eθ = ηβ2 I0l4π
[j
(βr)+
1(βr)2 − j
(βr)3
]e−jβrsinθ (6.8)
Eφ = 0 (6.9)
where the subscripts r, θ and φ, indicate the direction of the fields according to
spherical coordinate system; η is the wave impedance which equals 377 Ω in free
space; β is the propagation constant and is computed using λ2π ; I0 is the current
in the electric dipole; l is the length of the electric dipole; r is the distance from a
fixed reference point. It can be easily proven from (6.5) - (6.9), and by using the
formula for time-average power density:
Wav =12
Re[E × H∗] (6.10)
When r = λ2π , the first and the third terms in the bracket of (6.8) are the same in
magnitude but opposite in sign, and hence will cancel out each other. However,
when r < λ2π , the magnitude of the third term in the bracket of (6.8) will be greater
than the second term in the same bracket. In the case of r λ2π , the third term
in the bracket of (6.8) will completely dominate the other two terms. When this
happens, the energy storage field will be the dominant part while the power flow
(radiating field) will be negligible.
On the other hand, at a distance r > λ2π , the first terms of (6.6) and (6.8) will be
dominant. In this zone, power flow (radiating field) will be the dominant part
while the energy storage field will be negligible in comparison.
A similar inspection can be carried for a magnetic dipole. The E and the H fields
of a small loop or infinitesimal magnetic dipole can be expressed as:
Page 128
Chapter 6 RFID Tag Antenna Design
Er = Eθ = Hφ = 0 (6.11)
Eφ =β2 jωµ0 IA
4π
[j
βr+
1(βr)2
]e−jβrsinθ (6.12)
Hr =1η
β2 jωµ0 IA4π
[2
(βr)2 − 2j(βr)3
]e−jβrcosθ (6.13)
Hθ =1η
β2 jωµ0 IA4π
[j
βr+
1(βr)2 − j
(βr)3
]e−jβrsinθ (6.14)
where the subscripts r, θ and φ, indicates the direction of the fields according to
spherical coordinate system; η is the wave impedance which equals 377 Ω in free
space; β is the propagation constant and is computed using λ2π ; I is the current in
the magnetic dipole; A is the area enclosed by the magnetic dipole (loop); and r
is the distance from an origin at the centre of the loop.
When r λ2π , the third term in the bracket of (6.14) and the second term in the
bracket of (6.13) will completely dominate the other two terms (energy storage
field dominant), while at a distance r > λ2π , the first terms of (6.12) and (6.14) will
be dominant (radiating field dominant).
Hence, r = λ2π is defined as the boundary of the near field and far field. The near
field is the energy storing field while the far field is the energy radiating field. For
the operation of HF and UHF, this boundary is at a distance of 3.5 m and 0.052 m
respectively.
• Coupling and radiation
Coupling and radiation are two different means of power transfer from a point to
another point without using any physical connection between those two points.
In other words, power is transferred using a wireless connection; its possibility is
consistent with the Maxwell’s equations.
The coupling that normally takes place in the near field, which has been defined
previously, is in the form of either inductive or capacitive coupling. In such cou-
pling, energy flows out of the source in the form of electric or magnetic field and
flows back to the source. Ideally, there will be no energy loss if there is no resis-
tive load near the field. Hence, a near field is also known as the energy storage
Page 129
6.2 Antenna Theory
field. If there is a load in the near field, energy will be transferred from the source
to the load through coupling.
Radiation normally takes place in the far field, which has been defined previ-
ously. In the case of radiation, energy propagates or radiates away from the
source and never returns to the source. When a load adsorbs radiated energy,
the source will not be significantly affected.
However, the term “couple” or “coupling” is also commonly used in the litera-
ture with a different meaning. For example, the phrase “antenna A is coupled to
the magnetic field” does not mean that the antenna A is coupled in the near field.
It means that antenna A is sensitive to the magnetic field, whether it is in the near
or far field.
• Power Transfer
In the near field operation, to estimate the power received by an RFID tag chip,
coupling volume theory as introduced in [89] is applied.
Coupling volume is a figure of merit. For a magnetic dipole, the equations in-
volved are given by:
Vc =
Reactive power flowing in the untuned
label coil when it is short circuited
Volume density of reactive power
created by the interrogator at the label position
(6.15)
Vd =
Reactive power flowing in the inductor
of the interrogator field creation coil
Volume density of reactive power created
by the interrogator at the label position
(6.16)
If there are two magnetic dipoles resonating with quality factors Q1 and Q2 re-
spectively, the ratio between the power dissipated in the second magnetic dipole,
P2, and the power dissipated in the first magnetic dipole, P1, is given as:
Page 130
Chapter 6 RFID Tag Antenna Design
P2
P1=
Vc
VdQ1Q2 (6.17)
If the first magnetic dipole is an RFID reader antenna (transmitting antenna with
quality factor Q1), (6.17) can estimate the power received by an RFID tag, where
the RFID tag antenna is acting as the second magnetic dipole with quality factor
Q2.
The coupling volume for an electric dipole is slightly different in definition as
compared to the coupling volume for a magnetic dipole:
Vc =
Reactive power flowing in the untuned label capacitor
when it is open circuited
Volume density of reactive power
created by the interrogator at the label position
(6.18)
Vd =
Reactive power flowing in the capacitance
of the interrogator field creation electrodes
Volume density of reactive power created
by the interrogator at the label position
(6.19)
Similarly to the case of two magnetic dipoles, if there are two electric dipoles res-
onating with quality factors Q1 and Q2 respectively, the ratio between the power
dissipated in the second electric dipole, P2, and the power dissipated in the first
electric dipole, P1, is given as:
P2
P1=
Vc
VdQ1Q2 (6.20)
However, (6.17) and (6.20) are only meant for near field power transfer. In the far
field zone, radiation theory is often applied.
From radiation theory, the power transfer from the transmitting antenna to a loss-
less receiving antenna can be calculated using:
Page 131
6.2 Antenna Theory
Pr
Pt= gtgr
(λ
4πd
)2
(6.21)
where λ = wavelength; Pt = transmitted power; Pr = available received power;
gt = transmitter antenna gain; gr = receiver antenna gain; d = separation distance
between antennas.
It must be noted that (6.21) does not take in consideration of any mismatch be-
tween the antenna and the load. An RFID reader antenna often has a good match
with an RFID reader RF front end. However, owing to size constraints, small
RFID tags may not have a match between the RFID tag antenna and the RFID
tag chip. This mismatch loss must be taken into account using (6.22), and will be
discussed next.
• Maximum power transfer
A passive RFID tag normally does not have its own power source and it is pow-
ered by an interrogating antenna using CW transmission.
Hence to enable maximum read range, power transmission between the tag an-
tenna and the tag chip must be maximised. It can be proven easily that the maxi-
mum power transfer occurs when the impedance of the antenna is the conjugate
of the impedance of the tag chip. Note that the tag antenna is not a lossless an-
tenna.
If there is no conjugate match, the power loss can be computed through [85]:
PlostPavailable
=∣∣∣∣Zant − Z∗
cctZant + Zcct
∣∣∣∣2
(6.22)
where Zant is the input impedance of the antenna (in our case, Zant refers to the
designed RFID tag antenna) and Zcct is the input impedance of the circuit which
is connected to the antenna at its input terminals (in our case, this is the input
impedance of an RFID tag chip used in prototyping).
• Polarisation
Page 132
Chapter 6 RFID Tag Antenna Design
Polarisation is the curve traced by the end point of a vector representing the in-
stantaneous electric field [85]. In literature, there are three types of polarisation,
linear, circular and elliptical, though in fact a linear or a circular polarisation is a
special case of elliptical polarisation.
An RFID tag antenna is usually linearly polarised. However, the orientation of
an RFID tag antenna may change rapidly as the object it is attached to is being
transported along the supply chain. If a linearly polarised RFID reader antenna
is deployed and if the RFID tag antenna is aligned orthogonally to the reader
antenna, the reader will not be able to communicate with the RFID tag. Based on
this, RFID reader antennas are often circularly polarised.
Polarisation efficiency is defined as “the ratio of the power received by an an-
tenna from a given plane wave of arbitrary polarisation to the power that would
be received by the same antenna from a plane wave of the same power flux den-
sity and direction of propagation, whose state of polarisation has been adjusted
for a maximum received power” [85].
It can be observed easily, without mathematical proof, from a Poincare sphere
[86], that the polarisation efficiency between a circularly polarised RFID reader
antenna and a linearly polarised RFID tag antenna is 0.5 (or -3 dB).
• Scattering Parameters
Scattering parameters are mostly used in the study of distributed circuits involv-
ing transmission lines. In general, a scattering parameter represents the ratio
between an incident wave, a, to a port of interest, and an emerging wave, b, from
the same port or from another port. In particular,
Sij =bi
aj(6.23)
where aj is a known signal entering through port j; bi is the signal detected in
port i when port i is fed from a matched source or terminated with a matched de-
tector. It should be noted that all other ports (if there are any) must be terminated
with a matched load. The a and b parameters have the property that they are
Page 133
6.3 Antenna Parameters
in phase with the voltages (or transferred electric fields) of the travelling waves
and the amplitudes and units are adjusted so that the square of the magnitude
represents the power flow. In the usual case, the characteristic impedances, Z0,
of any transmission line involved are assumed to be of the same value, and the
scattering parameters can be defined as ratios of transmission line voltages.
In the measurement of the performance of a designed antenna, the antenna is
connected to a network analyser using a coaxial cable (a transmission line). A
normal antenna can be considered as a single port network. Hence, the most
common S parameter is S11. S11 is the ratio of the reflected signal to the incident
signal, also known as the input reflection coefficient, Γ.
Alternatively, input reflection coefficient can be computed using:
Γ =ZL − Z0
ZL + Z0(6.24)
where ZL is the input impedance of the antenna under test, and Z0 is the charac-
teristic impedance of the coaxial cable connected to the antenna. Z0 is normally
purely resistive and hence (6.24) is always in magnitude less than 1.
S21 is sometimes used when two antennas are involved (one as the transmitter
and the other one as the receiver). S21 describe the transmission quality between
those two antennas. In the case of two passive antennas, S21 = S12, due to reci-
procity.
In measuring an antenna using a network analyser, the reflection measurement
is defined as the ratio of reflected power to incident power, which is |S11|2.
6.3 Antenna Parameters
Common terms to describe the characteristics of an antenna, also known as the antenna
parameters, are referred to in various publications [85, 86, 90, 91, 92]. Definitions are
obtained from the “IEEE Standards Definitions of Terms for Antennas” [90] to avoid
any ambiguity in this thesis. Note that an antenna is assumed by Section 1.1 of [90] (not
Page 134
Chapter 6 RFID Tag Antenna Design
the Section 1.1 of this thesis) to be linearly polarised. This is not the case in RFID ap-
plications as an RFID reader antenna is usually circularly polarised. Special notes will
be added for any parameter which requires extra explanation for the case of circularly
polarised antenna.
1. Directivity
Directivity of an antenna is the ratio of the radiation intensity in a given direc-
tion from the antenna to the radiation intensity averaged over all directions. The
average radiation intensity is equal to the total power radiated by the antenna di-
vided by 4π. If the direction is not specified, the direction of maximum radiation
intensity is implied.
2. Gain
In simple form, the gain of an antenna is obtained through the reduction the di-
rectivity of an antenna by the amount equal to the dissipative loss of the antenna.
If an antenna does not have dissipative loss, the gain of this antenna is equivalent
to the directivity of this antenna.
Gain is defined as the ratio of the radiation intensity, in a given direction, to the
radiation intensity that would be obtained if the power accepted by the antenna
were radiated isotropically. Gain does not include losses arising from impedance
and polarisation mismatch.
Special notes: The gain of a linearly polarised antenna is quoted in “dBi”. In
the case of a circularly polarised antenna, antenna manufacturers will sometimes
quote the gain of an antenna using the term “dBic”, though some manufacturers
still prefer to use “dBi”. This may create some confusion. A simple rule is that
“dBic” of a circularly polarised antenna is 3 dB higher than an equivalent “dBi”
circularly polarised antenna.
3. Radiation Pattern
The radiation pattern is the spatial distribution of a quantity that characterises
the electromagnetic field generated by an antenna. The quantities that are most
Page 135
6.3 Antenna Parameters
often used to characterise the radiation from an antenna are proportional to, or
equal to, power flux density, radiation intensity, directivity, phase, polarisation,
and field strength. When the quantity is not specified, an amplitude or power
pattern is implied.
In this thesis, the radiation pattern plots show either the gain or directivity of the
antenna under test, unless otherwise stated.
4. Polarisation
See Section 6.2.
5. Input Impedance
The impedance presented by an antenna at its terminals. Impedance of an an-
tenna, Z, is normally presented in the form of R + jX, where R is the resistance
of the antenna and X is the reactance of the antenna.
6. Radiation Efficiency
The ratio of the total power radiated by an antenna to the net power accepted by
the antenna from the connected transmitter.
7. Quality Factor
Quality factor, Q, of a resonant antenna is the ratio of 2π times the energy stored
in the fields excited by the antenna to the energy radiated and dissipated per cy-
cle. For an electrically small antenna, it is numerically equal to one-half the mag-
nitude of the ratio of the incremental change in impedance to the corresponding
incremental change in frequency at resonance, divided by the ratio of the antenna
resistance to the resonant frequency.
8. Bandwidth
The range of frequencies within which the performance of the antenna, with re-
spect to some characteristic, conforms to a specified standard.
9. Radiation Resistance
Page 136
Chapter 6 RFID Tag Antenna Design
Radiation resistance is the ratio of the power radiated by an antenna to the square
of the r.m.s. antenna current referred to a specified point. The total power radi-
ated is equal to the power accepted by the antenna minus the power dissipated
in the antenna.
6.4 Resonant Circuit for Antenna
To estimate the performance of antenna, it is useful to visualise an antenna as its reso-
nant circuit, which consists of just resistors, inductors and capacitors. From a resonant
circuit, it is convenient to compute the quality factor of an antenna. However, it must
be stressed that this is purely an estimation as at high frequency, a distributed circuit
is more accurate than a discrete component resonant circuit. Nonetheless, a discrete
circuit is sufficient to provide a good initial estimation of the performance of a new
design.
There are three types of simple resonant circuits; series resonant circuits, parallel res-
onant circuits, and practical resonance circuit. The details of for these three resonant
circuits can be found in [93]. Some important results are shown here for the discussion
on how an RFID tag antenna or RFID tag chip can be represented by these equivalent
circuits.
• Series Resonant Circuit
A series resonant circuit consists of a resistor, Rs, an inductor, Ls, and a capacitor,
Cs, linked together in series as shown in Fig. 6.4.
R L
CV s
ss
Figure 6.4. Series resonant circuit.
The resonant frequency of this series resonant circuit is given by:
Page 137
6.4 Resonant Circuit for Antenna
f0 =1
2π√
LsCs(6.25)
The quality factor of this series resonant circuit is given by:
Q =XL
Rs(6.26)
where XL the reactance of the inductor at resonance.
The bandwidth of this series resonant circuit is given by:
BW =f0
Q(6.27)
where f0 is the resonance frequency of the circuit.
• Parallel Resonant Circuit
A parallel resonant circuit consists of a resistor, Rp, an inductor, Lp, and a capaci-
tor, Cp, aligned in parallel as shown in Fig. 6.5.
R L CI ppp
Figure 6.5. Parallel resonant circuit.
The resonant frequency of this parallel resonant circuit is given by:
f0 =1
2π√
LpCp(6.28)
The quality factor of this parallel resonant circuit is given by:
Q =Rp
XL(6.29)
The bandwidth of of this parallel resonant circuit is given by:
Page 138
Chapter 6 RFID Tag Antenna Design
BW =f0
Q(6.30)
• Practical Parallel Resonant Circuit
The parallel resonant circuit shown in Fig. 6.5 is often known as the ideal parallel
resonant circuit. In the case of antenna design, no actual resistor will be added in
the antenna. The resistor in a resonant circuit of an antenna normally represents
the losses introduced by the inductor. Hence, as an alternative, the circuit of
Fig. 6.6, also known a practical parallel resonant circuit, may be investigated.
r
L
CI p
p
Figure 6.6. Practical parallel resonant circuit.
Though Fig. 6.6 better characterises an antenna, it is easier to analyse mathe-
matically when an antenna is represented in the form of ideal parallel resonant
circuit. It can be done be transforming, approximately, the losses in inductor, r, to
the parallel resistor, Rp using:
Rp =(ω0Lp)2
r(6.31)
where ω0 is the angular resonant frequency of the circuit.
Different types of RFID tag chips are discussed in Section 6.6. For the case of HF chip,
the input impedance is just the input capacitance. Hence, the capacitor (Cs in Fig. 6.4
and Cp in Fig. 6.5) may be the input capacitance of the HF chip. The inductor and the
resistor represent the inductance of the HF coil antenna and the losses within the coil
respectively.
Page 139
6.5 Challenges in RFID Tag Antenna Design
The case of UHF RFID tag chips is more complex, in the sense that the input impedance
of a UHF RFID tag chip is represented by a resistor in parallel with a capacitor. For
example, a Texas Instruments RI-UHF-STRAP-08 UHF Gen2 Strap [94] (More details
in Section 6.6) has an input impedance specified by 380 Ω‖2.8 pF, or 9.7866 − j60.1926
in series.
A UHF RFID tag antenna normally has a resultant inductive input impedance to min-
imise the power transfer loss dictated by (6.22) between the antenna and its UHF RFID
tag chip. Hence, it can be represented by an inductor and resistor. Note that, this does
not mean an antenna does not have self-capacitance. It is just that at the frequency of
interest, the antenna is inductive in impedance.
A resonant circuit can then be formed using the inductor and resistor from the antenna,
together with the capacitor and resistor from the chip.
Note also that the bandwidth of a UHF RFID tag does not equal to the bandwidth of a
UHF antenna. This is caused by the additional resistance from the input impedance of
the UHF chip.
6.5 Challenges in RFID Tag Antenna Design
The design procedures for an HF RFID tag antenna are mature and well-documented,
such as in [95]. Basically, an HF RFID tag antenna provides enough inductance to
resonate with the HF RFID tag chip input capacitance, to operate at the frequency of
13.56 MHz. The HF tag is designed to have a reasonably high quality factor so as to
have a satisfactory read range (Some HF tags are purposely designed to have a low Q
as specified in 18000 Part 3 Mode 2). The read range of an HF system is often limited
by the reader antenna size, not the HF tag antenna, though in some cases the limiting
factor is the size of the tag antenna itself.
However it is not the same case for the design of UHF RFID tag antennas. Accord-
ing to [96], the main performance limitations of passive UHF RFID systems are chip
sensitivity, antenna gain, polarisation and impedance matching. All of the limitations
Page 140
Chapter 6 RFID Tag Antenna Design
mentioned above are related to an RFID tag antenna, apart from the limitation of chip
sensitivity.
The following investigation in this section details challenges in designing UHF RFID
tag antenna and is based on two co-authored papers: “Analysis of constraints in small
UHF RFID tag design” [97] and “Small UHF RFID label antenna design and limita-
tions” [98].
• Size Limitation
An electrically small antenna is often defined as an antenna with length less than1
10 of wavelength at its operating frequency.
For HF (13.56 MHz), 110 of wavelength is approximately 2.2 m while for UHF (860
- 960 MHz), 110 of wavelength is approximately 31 - 35 mm. Since the size of an
RFID tag always depends on the size of its antenna, and for most applications it
is not feasible to have a large RFID tag, an RFID tag antenna is often small in size
and considered as an electrically small antenna.
A small antenna often has small radiation resistance. From (6.26), a small r will
result in a high quality factor, Q. A high Q means small bandwidth.
Ref. [99] shows that many electrically small antennas can be matched to a 50 Ω
line. However, the quality factor of the antenna is said to be not significantly
changed.
• Bandwidth Limitation
According to Bode and Fano, a fundamental limitation on impedance matching
takes the form [100]:
∫ ∞
0ln
1|Γ|dω ≤ π
RC(6.32)
where Γ is the reflection coefficient of the load and its assumed lossless matching
network with respect to the source impedance RS, and R and C is the resistance
and capacitance, respectively, that comes from the parallel RC load.
Page 141
6.5 Challenges in RFID Tag Antenna Design
To have a wide matching frequency band, a maximum mismatch outside ∆ω (the
band of interest) is required.
The best utilisation of πRC is to keep |Γ| constant (say at |Γ|inband) over the band
∆ω, and unity outside this band as shown in Fig. 6.7.
Figure 6.7. Reflection coefficient for best utilisation of λRC .
Based on the case shown in Fig. 6.7, (6.32) becomes
|Γ|inband ≥ e−1
2∆ f RC (6.33)
It is found in practice that with the chip input resistances and capacitances nor-
mally encountered, good power transfer can be obtained over the bandwidth of
interest for RFID. More details can be found in [97].
• Gain Limitation
An RFID tag antenna is categorised as small antenna. An efficient small RFID tag
antenna is normally assumed to have a gain of 1.5. For an operating RFID reader
antenna the maximum transmitted power that is allowed under FCC regulations
is 1 W into an antenna with a maximum gain of 6 dBi.
• Polarisation Loss
Page 142
Chapter 6 RFID Tag Antenna Design
From Section 6.2, the polarisation efficiency between a circularly polarised RFID
reader antenna and a linearly polarised RFID tag antenna is 0.5 (or -3 dB). How-
ever, it is almost impossible to have a perfect circularly polarised antenna. An
antenna is more likely to be elliptically polarised and depending on the axial ra-
tio, the maximum polarisation loss will increase with respect to the -3 dB of a
circularly polarised antenna. Ref. [101] shows that with a 3 dB axial ratio (2:1
ratio), the maximum polarisation loss will increase to 4.77 dB.
• Impedance Matching
An RFID chip is always capacitive in nature due to its reservoir capacitor in its
rectifying circuit. A matching antenna to a capacitive RFID chip must be induc-
tive in nature.
In the case of HF, the inductance of the HF antenna is tuned to resonate with the
capacitance of the HF chip at the frequency of operation. The inductance needed
can be provided by a multi-turn loop antenna. More discussion can be found in
Section 6.10.
In the case of UHF, Eq. (6.22) is used to compute the power transfer from the
antenna to the chip. The best case is when the impedance of the UHF antenna is
the conjugate match of impedance of the UHF chip.
However, a simple dipole that is usually less than half-wavelength in length nor-
mally has a capacitive input impedance [85]. A matching network is required
to match the dipole to the RFID chip. Commercial UHF RFID tag antennas nor-
mally have a matching network very similar to a conventional T-match network,
which can be found in [85].
The discussions of impedance matching will be included in the presentation of
designed tag antennas throughout this chapter. An examination on the effect of
impedance mismatch on the maximum power transfer between a tag antenna
and a tag chip can be found in Section 9.6.
Page 143
6.6 RFID Chips
6.6 RFID Chips
This section describes the common RFID chips used in experiments to test the perfor-
mance of an RFID tag antenna in terms of readability and its maximum read range.
In the early stage, HF ICs were obtained from Texas Instruments (RF-HDT-SJME-G1
Tag-it HF-I transponder IC) [102] and NXP-Phillips (I Code SLI SL2ICS20 Smart Label)
[103]. Both of these tag ICs conform to ISO 18000-Part 3 Mode 1 [33] and ISO 15693 [34].
The problem is that there is no way to attach an IC directly on to an antenna. Hence
a discussion was held with a technical group from Defence Science and Technology
Organisation (DSTO), Australia. It was discovered that a solder bump can be used to
connect electrically an IC to a thin metal film as shown in Fig. 6.8.
HF Chip
Solder1.5 mm Bond
Metallic Pad Ceramic
3 mm
0.38 mm
Few micron
Solder
Figure 6.8. The proposed chip bonding on to a thin metal strip.
However, due to the cost factor, the idea of bonding of ICs on to a conductive strips
was abandoned. As an alternative, ICs from existing tags were reused. The tag of
choice is Texas Instruments Tag-It HF-I C04, as shown in Fig. 6.9(a). An HF chip is cut
out carefully from a Tag-It HF-I C04 with a small portion of the metal film remaining
attached. The chip is then attached on to a prototyped antenna as shown in Fig. 6.9(b).
There is an electrical connection from the chip to the antenna through the metal film.
A special tape is used to stick the metal film and the antenna together. A thorough
discussion can be found in Section 6.8.
The chip reused from a Tag-It HF-I C04 will be known as the TI HF chip in this the-
sis. It has an input capacitance of 23.5 pF ±10% [102] at the operating frequency of
13.56 MHz.
Page 144
Chapter 6 RFID Tag Antenna Design
(a) A Tag-It HF-I C04
Tag Chip
(b) The chip of a Tag-It HF-I C04
reused on prototyped antennas
Figure 6.9. Reuse of HF RFID tag chip.
Another tag chip which can be reused is the ISD72128 or C220, originally designed
by Integrated Silicon Design Pty Ltd (an Australian company) and manufactured by
Chartered Semiconductor Manufacturing (Singapore) as shown in Fig. 6.10. However
this tag chip uses a proprietary (Tag Talk First) protocol and can only be interrogated
by proprietary RFID readers, also available originally from Integrated Silicon Design
Pty Ltd. The functionality of this TTF chip is as follows with respect to the induced
voltage at its input terminals: (1) At low voltage, the tag chip will generate a 100 µs
burst of 100 kHz square wave. This is an EAS (Electronic Article Surveillance) burst. (2)
At higher voltages the tag will generate a reply burst, 4 cycles of either 250 or 400 kHz,
depending upon the code written to memory. The sequence repeats at a duty cycle
of between 10 to 20 percent, depending upon excitation levels, 10 percent for high
excitations and 20 percent for low excitations.
Figure 6.10. A ISD72128 or C220 HF RFID tag.
There is a variant of HF tag chips with much higher input capacitance. For example the
NXP-Phillips SL1 ICS31 01 ICODE1 Label IC [103] has an input capacitance of 97 pF.
The advantage of having a higher input capacitance is that the antenna for the chip
Page 145
6.6 RFID Chips
can have fewer turns to reduce the overall dimensions, as the antenna does not need
to provide a high inductance to resonate with the input capacitance of the chip.
Also used was the ISD72128 packaged in a smart card module. The tag chip is as
shown in Fig. 6.11. The tag chip which is already mounted on the smart card module
is easy to solder on to an antenna.
Figure 6.11. 72128 TTF chip on smart card module.
Unlike the difficulties found in mounting the HF chip on to an antenna substrate, UHF
chips are readily available attached by thin metal straps. A special tape (Section 6.8) is
used to attached a UHF strap on to a UHF antenna.
An example of a UHF strap used in the fabrication of a UHF tag is as shown in Fig. 6.12,
which is the RI-UHF-STRAP-08 UHF Gen2 Strap from Texas Instruments [94]. It has a
input impedance of 380 Ω‖2.8 pF which is 9.7866 − j60.193 Ω at 920 MHz.
Figure 6.12. A Texas Instruments UHF strap.
Another type of UHF strap used in experiments is the C1G2 UHF strap from Alien
Technology (Fig. 6.13). It has an input impedance of 1.5k Ω‖1.3 pF which is 11.713 −j132.03 Ω at 920 MHz.
Page 146
Chapter 6 RFID Tag Antenna Design
Figure 6.13. An Alien C1G2 UHF strap.
6.7 RFID Readers
This section lists the common RFID readers used for the all the testing presented in
this thesis. HF readers operate at 13.56 MHz. UHF readers can be controlled to op-
erate within 860 to 960 MHz. Unless testing was carried in a controlled environment,
such as an anechoic chamber, the measurement results presented in this thesis for UHF
were carried out in adherence to Australian regulations, that is in the band of 920 to
926 MHz, at 4 W EIRP (Effective Isotropic Radiated Power).
6.7.1 HF Reader
• Fig. 6.14 shows an ID ISC.LR2000 HF reader and an ID ISC.ANT300 HF reader
antenna, both from FEIG Electronics. It supports ISO 15693, ISO 18000 Part 3
Mode 1.
• Fig. 6.15 shows an L120 HF reader and an HF AC201 reader antenna, both from
Gemplus Tag Australia, previously known as Integrated Silicon Design Pty Ltd.
It supports the proprietary TTF protocol. The suitable RFID tags are the ISD72128
(C220).
6.7.2 UHF Reader
• Fig. 6.16 shows an ID ISC.LRU2000 UHF RFID reader and an ID ISC.ANT.U250/250
UHF reader antenna, both from FEIG Electronics. It supports 18000 Part 6, EPC
C1G1 and EPC C1G2 tags.
Page 147
6.7 RFID Readers
1
23
(a) Reader
1
(b) Antenna
Figure 6.14. FEIG HF RFID Reader. (a) 1: Connection to reader antenna; 2: Optional connection
to a computer through RS232 interface; 3: Connection to a computer through a LAN
connection using TCP/IP protocol. (Connection to power supply not shown) (b) 1:
Connection to a reader.
123
(a) Reader
1
(b) Antenna
Figure 6.15. Gemplus HF RFID reader. (a) 1: Connection to reader antenna; 2: Connection to
a computer; 3: Connection to power supply. (b) 1: Connection to a reader. Note that
there is a ferrite core on the cable to reduce noise.
1
2
(a) Reader
1
(b) Antenna
Figure 6.16. FEIG UHF RFID reader. (a) 1: Connection to reader antenna; 2: Connection to a
computer. (b) 1: Connection to a reader.
Page 148
Chapter 6 RFID Tag Antenna Design
• Fig. 6.17 shows an ALR 9780 UHF RFID reader and an ALR 9610-BC reader an-
tenna, both from Alien Technology. It supports EPC C1G1 and EPC C1G2.
1
2
3
4
(a) Reader
3
(b) Antenna
Figure 6.17. Alien UHF RFID reader. (a) 1: Connection to a computer using serial connection; 2:
Connection to power supply; 3: Connection to reader antennas; 4: Unused connection
to antennas. (b) 3: Connection to a reader.
6.8 RFID Tag Fabrication
An RFID tag is the combination of an RFID tag antenna and an RFID tag chip, where
the tag chip is located across the input terminal of the RFID tag antenna.
For prototyping, the raw material for an RFID tag antenna is the Printed Circuit Board
(PCB). An example is the commonly used FR4 board. There are many types of PCB
boards uasable in the prototyping of the design RFID tag antenna. For the conductive
layer, copper and aluminium are the most common. For the dielectric layer, FR4 is the
most common.
There is a selection of choices for the thickness of the dielectric layer. The commonly
used ones are 1.6 mm and 0.2 mm. Also, the thickness of the metallic layer can be
chosen from within a few common values.
Most of the time, double sided PCB boards are needed for most of the antenna designs.
For the HF coil antenna, the bottom layer is needed for the underpass, which is linked
to the top layer through via holes.
Via holes are created using a rivet if a pad size of at least 1.3 mm in diameter is possible.
If, due to space constraints and no pad with at least 1.3 mm being available, a via hole
Page 149
6.9 Antenna Design and Simulation
is created by first drilling a hole using mini drill (Smallest possible diameter is 0.3 mm)
and a fine conductor is passed through the hole and soldered at both ends.
For the case of the UHF antenna, the bottom layer is always used for matching pur-
poses. The most common method is to have the bottom layer providing series or paral-
lel capacitance. No physical connection is needed between the top and bottom layers.
A typical HF coil antenna, with low fabrication cost, is a planar spiral strip antenna
on a dielectric substrate, as shown in Fig. 6.18. The dielectric substrate chosen in this
research work when fabricating an HF antenna is FR4, with relative permittivity of 4.4.
Figure 6.18. Typical HF antenna. Planar spiral in shape on a dielectric slab, normally FR4.
To attach a tag chip, 3MTM Z-Axis Electrically Conductive Tape 9703 is used. As sug-
gested by its name, this conductive tape only conducts in the Z-axis. It has a low
contact resistance . However it is not clear, from the specification sheet, what capac-
itance it may introduce into the circuit. A simple experiment was carried out to test
the effect of this conductive tape on an RFID tag. As shown in Fig. 6.19, a small part
of the HF coil antenna was removed. A small copper tape was then attached using
the Z-axis conductive tape to replace the removed HF coil track. A measurement us-
ing a Gemplus L120 reader shows that there is no degradation in performance. Hence
it is concluded that the Z-axis conductive tape has minimal impact on the RFID tag
antenna.
6.9 Antenna Design and Simulation
Page 150
Chapter 6 RFID Tag Antenna Design
Figure 6.19. Effect of using Z-axis conductive tape on an HF RFID tag.
6.9.1 Ansoft HFSS
High Frequency Structure Simulator (HFSS) is a 3D electromagnetic-field simulator
from Ansoft. This section documents the setup undertaken for all the simulations car-
ried out for RFID antenna design, for both the HF and UHF cases. This section is based
on the HFSS User Manual from Ansoft [104].
Ansoft HFSS uses the finite element method (FEM) to solve the simulation. In the
literature, there are other types of solver, which include the popular finite-difference
time-domain (FDTD) and the method of moments (MOM). This thesis does not inves-
tigate the difference between these solvers. A good description of the different solvers
or computational electromagnetic techniques can be found in [105].
Below are the common HFSS settings for all the simulations carried out for antennas
presented in this thesis:
Solution Type:
• Driven Modal Solution
For the calculation of modal-based S-parameters of passive, high frequency struc-
ture (including microstrip, transmission line). The scattering parameter solutions
will be expressed in terms of the incident and reflected powers of waveguide
modes.
Page 151
6.9 Antenna Design and Simulation
• Driven Terminal Solution
For the calculation of terminal-based S-parameters of multi-conductor transmis-
sion line ports. The scattering parameter solutions will be expressed in terms of
terminal voltages and currents.
• Eigenmode Solution
For the calculation of eigenmodes and resonances of a structure.
From experience, for the purpose of antenna simulation, the Driven Modal Solution is
the most appropriate choice. The Driven Terminal Solution is more suitable for circuit
analysis.
Excitation Point
Excitations in HFSS are used to specify the sources of electromagnetic fields and charges,
currents, or voltages on objects or surfaces in the design.
There are several types of excitation points in HFSS. The definition given by Ansoft for
the two common types are listed below:
1. Wave Port
Represents the surface through which a signal enters or exits the geometry.
2. Lumped Port
Represents an internal surface through which a signal enters or exits the geome-
try.
The difference between a wave port and lumped port is subtle. Both are represented by
a geometrical square or rectangle shape as an opening, which allows a signal to enter
or exit a structure (antenna in our case).
As suggested by the name, a wave port is often used to represent a waveguide or a
transmission line and is characterised by a characteristic impedance. In simulation a
wave port has to be defined at the edge of a structure.
Page 152
Chapter 6 RFID Tag Antenna Design
On the other hand, a lumped port can be characterised by a user defined complex
impedance and can be located internally. It is normally used for a microstrip structure.
Almost all of the antennas designed and presented in this thesis are fabricated on PCB
boards and have good resemblance to a microstrip structure. Hence, a lumped port is
chosen. The impedance of lumped port is set as 50 Ω instead of the impedance of the
RFID chip used.
A quick test showed that a change in the impedance of the lumped port will affect the
simulation results. For example, the input impedance of the antenna will change from
20 + j170 Ω at 950 MHz to 25 + j182 Ω at 950 MHz if the lumped port impedance is
changed from 15 Ω to 50 Ω. The reason a 50 Ω lumped port is chosen for simulation
is that the network analyser used for input impedance measurement is a 50 Ω system.
Hence, choosing 50 Ω will make the comparison between measurement results and
simulation results more meaningful and more accurate. More in-depth discussion on
measurement techniques and challenges is presented in Chapter 9.
Also, experience shows that simulation using a capacitive lumped port (for example
5 − j150 Ω to represent a capacitive UHF RFID tag) will result in much longer simula-
tion times.
The disadvantage of using a 50 Ω lumped port is that simulation results based on a
50 Ω lumped port may not be reflecting the actual situation when a UHF RFID tag is
connected to the antenna. To overcome this short-coming, most RFID tag antennas are
designed to have a structure with tuning flexibility. Since the difference in impedance
is not significant, a quick fine-tuning can optimise the performance of a RFID tag.
Boundary Condition
Similar to excitation types, there are several options for setting up a boundary for a
simulation. A boundary is required to specify the region of the problem. Without a
boundary, HFSS would not be able to confine the problem to a reasonable size.
The two common types of boundary for antenna simulation are as follow:
1. Radiation Boundary
Page 153
6.9 Antenna Design and Simulation
In HFSS, radiation boundaries are used to simulate open problems that allow
waves to radiate infinitely far into space, such as antenna designs. HFSS ab-
sorbs the wave at the radiation boundary, essentially ballooning the boundary
infinitely far away from the structure.
2. Perfectly Matched Layers (PML)
Perfectly matched layers (PML) are fictitious materials that fully absorb the elec-
tromagnetic fields impinging upon them. These materials are complex anisotropic.
Figure 6.20. PML layer in HFSS. The larger cubicle in the middle defines the working zone, and
is surrounded by PML layers.
In HFSS the radiation boundary must be at least quarter of wavelength away from the
radiating structure of interest. Also, a radiation boundary must be convex with respect
to a radiating source. This poses a lot of difficulties for the simulations of antennas
presented in this thesis. For example, at HF (13.56 MHz), a quarter of wavelength
is approximately 5.53 m. The antenna size is often in the size of up to 0.1 m. This
means in the simulation, the antenna must be located in a huge empty region, and the
simulation duration will increase significantly.
One may argue that an RFID HF antenna operates through inductive coupling and not
radiation. Hence a small radiation boundary is sufficient. However, HFSS computes
the power entering an antenna of interest and the power reflected from the antenna,
in order to compute the impedance of the antenna. With a radiation boundary smaller
than the recommended quarter wavelength, more power will be reflected, reducing
the accuracy of the simulation.
Page 154
Chapter 6 RFID Tag Antenna Design
One can compare a simulation to an actual measurement. Having a radiation boundary
in simulation is comparable to having absorbing material in an anechoic chamber for a
measurement. The absorbing material must be a certain distance away depending on
the frequency of operation of the antenna under test.
Also, the radiation boundary must be convex in shape with respect to the radiating
source. It is easy to fulfill this requirement if the antenna under test is a simple loop
antenna or a simple dipole. For a more complex antenna with an unsymmetrical or
irregular shape, it is very difficult to construct an appropriate radiation boundary.
6.9.2 Scripting in HFSS
Although the HFSS working environment allows 3-D modelling and HFSS supports
parametric optimisation, to construct spiral planar coil structure of many turns (> 4
turns) using HFSS is very cumbersome. The latest version of HFSS (version 10) sup-
ports modelling using a mathematical equation. However, an HF coil antenna is not
readily defined mathematically.
The alternative is to use the scripting function in HFSS to create the HF coil. The com-
prehensive HFSS guide is provided by Ansoft [106] and Virtual Basic Script (VBS) is
used as the programming language in the scripting of HFSS.
However, VBS lacks analytical tools. To write a VBS programme from scratch is time
consuming. Hence, it was decided to use MATLAB as the analytical tool. MATLAB
and HFSS are linked together to produce the desirable results. MATLAB will compute
the geometry of the structure using mathematical equations and compile HFSS script
using Virtual Basic Script (VBS). The link between MATLAB and HFSS is shown in
Fig. 6.21.
The foundation of the work is largely based on the API produced by V. C. Ramasami
[107]. Extensive extensions and modifications of the API have been made to suit the
work of the author.
The complete MATLAB script can be found in Appendix D. This MATLAB script de-
fines the function “makecoil”, which takes in the following parameters:
Page 155
6.9 Antenna Design and Simulation
MATLABScript
VB ScriptHFSS
Simulation
generate run
export result
Figure 6.21. Linking MATLAB and HFSS. MATLAB script is written to generate VB script. The
VB Script is then fed into the HFSS simulator to construct the antenna model and
set up simulation parameters. The simulated results are then exported from HFSS to
MATLAB for analysis and plotting.
• turn: This is to specify the number of turns of the coil antenna.
• L: The length of the inner-most coil.
• L1: The width of the inner-most coil.
• thick: The thickness of the copper film on top (or other conductive material). It is
normally set as 0.018 m.
• height: The thickness of the dielectric board. Normally a FR4 board of 1.6 mm
thickness is used.
• w: The width of the HF coil track.
• g: The gap between the coil tracks. A gap of 0 mm will result in a rectangular
patch as all the coil tracks will be merged together.
• save loc: The location and the filename of simulated results. The files will contain
both the S11 and the input impedance of the simulated antenna with respect to a
default frequency range of 860 to 960 MHz.
An example of an automatic generated HF coil model is as shown in Fig. 6.22.
A comparison between Table 6.2 (results from HFSS simulation) and Table 6.3 (results
from Section 6.10.2 based on self and mutual inductance concepts) and shows that
when the number of turns is small (< 6 turns), the simulated results are very close
to the computed results. However, as the number of turns increases, the difference
between these two results increases.
Page 156
Chapter 6 RFID Tag Antenna Design
(a) 2-loop Coil (b) 10-loop Coil
Figure 6.22. The auto generated loop using VBS.
Table 6.2. Inductance values for spiral planar coil with different numbers of turns, with L =
20; L1=40; w=3; g=1. The simulation of reactance considers both the inductance and
capacitance of a coil antenna. An effective inductance is obtained though Z = 2π f L.
Turns Reactance at 13.56 MHz (Ω) L (µH)
1 13.7120 0.16106
2 50.0168 0.58748
3 108.090 1.26950
4 191.700 2.25170
5 315.140 3.70160
6 504.470 5.92540
7 772.480 9.07340
8 1203.50 14.1360
9 2155.40 25.3170
10 4351.00 51.1060
In simulation, the impedance of the antenna is computed, not the inductance value. As
an approximation, the impedance is transformed back to the inductance value, using
Z = 2π f L.
6.9.3 Plotting in HFSS
HFSS can provides results in several forms. The most common results obtained are the
input impedances of the simulated results.
Page 157
6.9 Antenna Design and Simulation
In the simulation of an HF RFID antenna, the resonant frequency of an antenna can be
observed from the plot of real and imaginary parts of the simulated impedances. At
resonance the real part of the impedance will increase towards infinity and the imagi-
nary part will be zero.
In the simulation of the UHF RFID antenna, the input impedance can be presented
in the plot of impedance versus frequency on a Smith Chart. Also available from the
simulation of the UHF RFID antenna is the bandwidth of the antenna. To obtain the
bandwidth of the antenna itself is easy, by using a 50 Ω lumped port for simulation,
and plotting the |S11| of the simulation results. However, this |S11|, as shown in Trace
1 in Fig. 6.23 shows the bandwidth of the combination of the tag antenna and the
50 Ω source, but does not offer any insight into the bandwidth of the RFID tag alone.
Moreover, the bandwidth of an RFID tag is affected by both the tag antenna and the
chip attached to the antenna.
0.85 0.90 0.95 1.00 1.05-25.00
-20.00
-15.00
-10.00
-5.00
0.00
Freq [GHz]
1
2
Figure 6.23. Plots |S11| curve in HFSS. (Y-axis in dB). Trace 1: |S11| from the simulation which
represents the |S11| which would be seen if the antenna is connected to a measuring
instrument using a 50 Ω cable; Trace 2: Eq. (6.22) in dB terms, representing the
suggested way to plot a |S11| curve in HFSS, with an assumed chip impedance.
Normalising the simulation results using the target impedance of the designed antenna
is one way suggested by HFSS to obtain information on tag bandwidth. HFSS has a
built in post processing function to normalise results based on a new port impedance
Page 158
Chapter 6 RFID Tag Antenna Design
value. The complete arithmetic for the normalising process can be found in [104]. In
simple term, HFSS computes a new |S11|, or |Snew| using:
|Snew| =
∣∣∣∣∣Z0
1+S111−S11
− Znew
Z01+S111−S11
+ Znew
∣∣∣∣∣ (6.34)
where Znew is the target impedance of the designed antenna (or in other words, the
conjugate impedance of the chosen RFID chip.)
An example is shown in Fig. 6.24. Trace 1 is showing the plot of 20 log10 |Snew|, which
is obtained after the post processing normalisation.
0.85 0.90 0.95 1.00 1.05-40.00
-30.00
-20.00
-10.00
0.00
Freq [GHz]
1
2
Figure 6.24. Plots in HFSS showing the normalised |S11| curve. (Y-axis in dB). Trace 1:
Normalised |S11|; Trace 2: Eq. (6.22) in dB terms.
However, a problem lies in the fact that normalisation only works sensibly when the
new target impedance is purely real, which is not the case here as the target impedance
(a chip impedance), is both resistive and capacitive. As can be seen from Fig. 6.24, the
arithmetic of normalisation results in very low |Snew| value and it is not possible to
determine the bandwidth of the RFID tag. Also, this method assumes a constant RFID
chip input impedance, which in fact is frequency dependent.
The solution is to use (6.22) as the foundation in determining the bandwidth of an RFID
tag. Firstly, from Section 6.6, an equivalent circuit for a UHF RFID tag chip consists of a
Page 159
6.10 HF RFID Antenna Design
resistor in parallel with a capacitance (in the example shown in this section, the resistor
is assumed to be 1 kΩ and the capacitor is assumed to be 1 pF). Using this equivalent
circuit, the resultant input impedance of a UHF RFID tag chip within a frequency band
can be computed. Note that the input impedance changes as the frequency increases.
By substituting this computed input impedance of the chip and also the simulated
antenna impedance from HFSS into (6.22), the ratio between the power loss and the
power available can be computed. This ratio plotted in dB and is shown as Trace 2 in
both Fig. 6.23 and Fig. 6.24.
The bandwidth of an RFID tag (tag antenna and tag chip) can then be obtained from
Trace 2. Generally, the bandwidth is defined as the frequency range where the |S11|curve is below the -3 dB line. Within this bandwidth, an RFID tag is expected to func-
tion properly according to its specification.
6.9.4 ISO-Pro
ISO-Pro is a software program from T-Tech [108] to control the T-Tech’s quick proto-
typing system.
T-Tech’s quick prototyping system is actually a milling machine. It mills away un-
wanted metal from a PCB, and the remaining metal forms the designed antenna.
The 3-D antenna model is drawn in HFSS for simulation. The same simulation model is
used to redraw the designed antenna in Protel (Protel is a PCB layout software, which
can export a designed layout to ISO-Pro with ease).
From HFSS, the antenna model is exported in the format of AutoCAD DXF format.
However, a DXF format is a 2-D format. Hence, if a model has several layers, several
DXF files are needed.
6.10 HF RFID Antenna Design
This section presents the design of an HF RFID antenna. An HF RFID system operates
at 13.56 MHz. An HF RFID antenna is designed so that together with an RFID tag
Page 160
Chapter 6 RFID Tag Antenna Design
chip, it resonates at the operating frequency. Firstly, this section discusses the use of
a simple loop antenna as an HF RFID antenna. In the later part, various novel HF
antenna designs for various applications are shown.
The literature, [95, 109, 110, 111] provide guidelines in designing HF RFID antennas.
The fundamental idea in designing an HF antenna is always to have the designed an-
tenna providing sufficient inductance to tune with the HF chip capacitance and hence
to have a resonant point at the operating frequency. This section is not meant to over-
lap with existing literature. This section touches lightly on the theory and focuses
more on the application-driven engineering challenges: from understanding a unique
RFID application, initial antenna simulation, to antenna prototyping and finally an-
tenna testing.
6.10.1 A Simple Loop Antenna
A simple loop antenna acts as a magnetic dipole. The formulae used and the detailed
design procedures are described by Balanis [87] and Bernhard [92]. Optimisation of an
HF antenna is discussed by Chen [112] while Lee offers a design note on RFID antenna
circuit design [109].
An HF electrically small circular loop antenna with radius a and wire radius b [87] has
a radiation resistance given by:
Rr = 20π2(βa)4 Ω (6.35)
and inductance,
L = µoa[
ln(
8ab− 2
)](6.36)
For a square or rectangular (not circular) loop antenna, its characteristics can be esti-
mated using the area encompassed by the loop:
πr2 = a × b (6.37)
Page 161
6.10 HF RFID Antenna Design
where r is the radius of an equivalent circular loop antenna, a and b are the width and
the length of the rectangular or square (where a = b) respectively.
6.10.2 A HF Planar Spiral Coil Antenna
The characteristics of a simple one-turn planar coil antenna can be estimated using the
characteristics of a simple loop antenna. The equivalent of a simple one-turn planar
coil antenna as a simple loop antenna can be found by using the table of conductor
geometrical shapes and their equivalent circular cylinder radii [85]. A width of the
planar coil, a, is related to the radius of the cylindrical coil, ae, though the relation
ae = 0.25a.
For a multi-turn coil antenna where the coil maintains a fixed radius, the total induc-
tance can be estimated using:
Ltotal = n2 × L (6.38)
where Ltotal is the total inductance, L is the inductance of one loop, and n is the number
of loops.
However, it is impossible to have a multi-turn planar coil antenna with fixed radius.
Hence to predict the total inductance of a multi-turn planar coil antenna, a more com-
plex method from [113] is applied. (6.39) to (6.45) and their explanations are adapted
from [113].
This complex method is based on the idea of inductance per unit length. However, in
a physical world, the idea of inductance per unit length does not make much sense as
this idea implies that a short straight wire segment has a certain amount of inductance.
Hence, any statement about the inductance per unit length or the inductance of a wire
segment must always be recognised as being only a contribution to the true inductance
(self or mutual) of a complete circuit(s) (in our case is a loop antenna consists of more
than 4 wire segments) and that the complete inductance will consist of the sum of the
so-called inductances of all wire segments and so-called mutual inductances between
all segments. A comprehensive discussion on this matter can be found in Appendix B.
Page 162
Chapter 6 RFID Tag Antenna Design
Bearing that in mind, the exact expression for the inductance of a straight conductor
(or wire segment) is given as:
L = 0.002l[
ln(
2lGMD
)− 1.25 +
AMDl
+(µ
4
)T]
(6.39)
where
• L = inductance (µH)
• l = conductor length (cm)
• GMD = Geometric Mean Distance of the conductor cross section. GMD is defined
as the distance between two infinite thin imaginary filaments whose mutual in-
ductance is equal to the mutual inductance between the two original conductors.
For a single conductor, GMD is defined as the distance between two imaginary
filaments normal to the cross section, whose mutual inductance is equal to the
self-inductance of the conductor. The computation of GMD is often lengthy but
for common cross sectional shapes, the GMD can be obtained from the literature.
In the case of planar coil, where the cross sectional shape is a very thin rectangle,
the value for GMD is given by 0.22313× (a + b), where a and b are the width and
height of the cross sectional rectangle.
• AMD = Arithmetic Mean Distance of the conductor cross section. AMD for a
single conductor is the average of all possible distances within the cross section.
In the case of planar coil, where the cross sectional shape is a very thin rectangle,
the values for AMD is given as 13 .
• µ = conductor relative permeability.
• T = frequency-correction parameter. Although T can be considered 1 for mi-
crowave frequencies, we still assumed a value of 1 in HF. The argument for that
is out of the study of this research and is an assumed value.
To reduce (6.39) for our inductance calculation, we use T = 1, µ = 1, and consider the
copper strips (of the antenna loop) as having a very thin rectangular cross sectional
shape (i.e. b → 0):
Page 163
6.10 HF RFID Antenna Design
L = 0.002l
ln[
2la + b
]+ 0.50049 +
[(a + b)
3l
](6.40)
The aim is to compute the inductance of a planar spiral antenna. The total inductance
is contributed by both self inductance and mutual inductance by all the segments of
the coil. The self inductance can be obtained by calculating the self inductance of each
segment using (6.40) and summing all the self inductances of the segments.
The computation of mutual inductance is more complex, as there are positive and neg-
ative mutual inductances. Positive mutual inductance is caused by two parallel con-
ducting segments with the same current direction while negative mutual inductance
is caused by two parallel conduction segments with different current directions. We
assume that there is negligible mutual inductance between two perpendicular con-
ducting segments.
Hence the total inductance of a planar coil antenna is given by (6.41):
LTotal = Lsel f + M+ − M− (6.41)
where Lsel f is the total self inductance, M+ is the total positive mutual inductance of
all segments (in nH), M− is the total negative mutual inductance of all segments (in
nH).
Between a pair of parallel conducting segment, there exists either a positive or negative
mutual inductance. The magnitude for both positive or mutual inductance can be
calculated using (6.42):
M = 2lQ (6.42)
where l is the length of the segment (in cm), and Q (do not confuse with quality factor,
Q) is the mutual inductance parameter calculated using (6.43).
Q = ln
⎧⎨⎩
(l
GMD
)+
[1 +
(l
GMD
)2] 1
2⎫⎬⎭ −
[1 +
(GMD
l
)2] 1
2
+(
GMDl
)(6.43)
Page 164
Chapter 6 RFID Tag Antenna Design
However, if the two parallel conducting segments are not the same in length (which is
the case of a planar spiral antenna), the mutual inductance between them has to be
calculated using (6.44) and this equation is best explained by Fig. 6.25.
2MA,B = (Mb+p + Mb+q) - (Mp + Mq) (6.44)
where MA,B represents the mutual inductance between segment A and segment B, and
Mlength (such as Mb+p or Mp) can be computed using (6.42) and (6.43).
NOTE: This figure is included on page 165 of the print copy of the
thesis held in the University of Adelaide Library.
Figure 6.25. Computation of the mutual inductance between two planar wire
segments.
In the figure, planar wire segment A is j in length while planar wire segment B is m in
length. The difference length between planar wire segment A and B is (p+q). (Based
on [113].)
In the computation of the inductance of the planar spiral antenna, it is assumed that
the length of p and q are the same and equation (6.44) is then reduced to:
MA,B = Mb+p - Mp (6.45)
This planar spiral antenna must provide sufficient inductance so that the reactance
provided by the inductance can cancel the reactance provided by the chip capacitance
at the frequency of interest (which is 13.56 MHz for HF operation). If we assume a
chip capacitance of approximately 1 pF in series, we would need a series inductance
of 140 µH in series, calculated using:
(6.46)
Page 165
6.10 HF RFID Antenna Design
This amount of inductance is difficult to achieve using a spiral planar coil structure.
The normal practice is to increase the capacitance instead, by attaching a parallel ca-
pacitor to the chip’s input capacitance. However, at this stage, it is not known what
the impact of the additional parallel capacitance to the self-resonance frequency of the
spiral planar coil will be.
The computation of positive and negative mutual inductances is complex and lengthy.
Hence a MATLAB script is written to compute these mutual inductances. The complete
code is attached as Appendix E.
Together with the self inductance (refer to (6.40)), the total inductance can be com-
puted. Table 6.3 shows some of the results for spiral planar coil on different dimen-
sions:
Table 6.3. Inductance values for various spiral planar coils. The length of the inner-most coil
= 20 mm; the width of the inner-most coil = 40 mm; the width of the track = 3 mm;
the gap between the coil tracks = 1 mm. Lsel f is the self inductance of spiral coils; M−
is the total negative mutual inductance; M+ is the total positive inductance; LTotal is
the total inductance of spiral coils.
Turns Lsel f M− M+ LTotal (µH)
1 0.2093 0.0333 0.0352 0.2112
2 0.4631 0.1413 0.3177 0.6395
3 0.7513 0.3391 0.8781 1.2903
4 1.0747 0.6426 1.7247 2.1568
5 1.4339 1.0683 2.8746 3.2402
6 1.8294 1.6330 4.3485 4.5448
7 2.2617 2.3543 6.1691 6.0765
8 2.7313 3.2499 8.3604 7.8419
9 3.2387 4.3381 10.947 9.8476
10 3.7842 5.6377 13.954 12.101
Page 166
Chapter 6 RFID Tag Antenna Design
6.10.3 An HF Antenna for a New Wine Closures
This section describes the design process for an RFID HF tag antenna for a special wine
closure. This wine closure is a new invention from Zork [114]. It is made of plastic
with a thin metal foil on top to optimise its functionality as a wine closure, which is
to preserve the quality of the wine. Fig. 6.26 shows a model of the wine closure of
interest. An RFID tag is confined in the space shown in the figure.
Plastic Cover
Thin Metal Sheet
Space for TagPlastic Stopper
Figure 6.26. A Zork wine closure model. This diagram is not to scale and does not represent
an actual Zork wine closure. This diagram is to indicate the location of the space
available for RFID tag.
Figure 6.27. RFID tag for wine closure.
The aims of embedding an RFID tag into the wine closure from the point of wine re-
tailer include product authentication, tracking, and anti-theft and anti-counterfeiting.
From the view of customers, they can scan a wine bottle embedded with an RFID tag
in some specific area to obtain extra information about the wine they have just chosen,
Page 167
6.10 HF RFID Antenna Design
such as the the origin of the wine, ingredients, awards, and even recommendations
from wine critics around the world.
The aim is to have a multi-turn coil antenna to provide inductance to resonate with
the HF chip input capacitance. Besides the space limitations, it is expected that the
thin metal foil will reduce the performance of the RFID tag. The final design of the HF
tag antenna is shown in Fig. 6.27. Table 6.4 shows the details and read range perfor-
mance of the designed HF RFID tag for Zork wine closure. It should be noted that the
maximum read Range is limited by the reader antenna size [95].
Table 6.4. Performance of HF RFID tag for Zork.
Material Copper wire
Dimension (mm) 25.0 (h) 7.5 (r)
Chip ISD72128 or C220
Reader ISD L120
Read Range (mm) 150 (Best)
6.10.4 An HF Antenna for a Pigs
This section describes the design process for an RFID HF tag antenna for a pig ear tag.
The pig ear tag project is sponsored by the Pork CRC, Australia. The design of a UHF
RFID tag antenna for pig ear tagging is presented in Section 6.11.4. The pig ear tag
is a relatively large project, when compared with the Zork wine closure tag antenna
design.
The common frequency used in livestock tagging is the Low Frequency (LF) band ,
spanning 125 - 135 kHz. LF RFID operation is standardised by ISO 11784/85 [51, 52].
The LF RFID system normally cannot handle a dense tag environment. Also, LF re-
quires large antenna components and hence is difficult to implement and is susceptible
to electrical noise, which HF can handle [20].
Page 168
Chapter 6 RFID Tag Antenna Design
Although a new enhanced version of an LF chip is available [115] with anti-collision
capability, the traditional LF tag which complies with ISO 11784/85 does not offer such
anti-collision capability. Without anti-collision, a single LF tag must only be present in
the reader antenna interrogation zone at any one time. For practical real life applica-
tions, animals are forced to pass through (or enter the interrogation zone) one by one.
Where this is practicable is when identifying larger animals, such as cattle or sheep.
Examples include cattle and sheep farming.
However, in pig farming, a problem arises. Small baby piglets are smaller in size as
compared with cattle. It is not easy to herd piglets through a gate mounted with an
antenna. Hence it is the intention of this research work to investigate the feasibility of
using either an HF or UHF RFID tag (Section 6.11.4) for the tagging of pigs, by using
the development of pig ear tag as the case study.
Livestock ear tags for cattle deployed in Australia must be accredited by National Live-
stock Identification System (NLIS), Australia. Examples of these livestock ear tags with
embedded RFID tag can be found in [116, 117], which is ISO 11784/85 compatible.
These tags can be recovered at slaughterhouse and reused, and can withstand haz-
ardous environments. A cattle livestock ear tag before any RFID tag is embedded
is shown in Fig. 6.28(a). One part of the ear tag has a hole in the middle to allow
both parts to be pinned together on the ear of cattle livestock. The space to attach
an RFID tag is hence restricted to a circular disc with a hole in the middle as shown
in Fig. 6.28(b). The outer diameter of the space available is 28.3 mm, while the inner
diameter is 12.0 mm.
In the early design stage, antennas were designed to be fitted into the casing shown in
Fig. 6.28. In the later stage of the research, the new industrial partner has provided a
new casing for encapsulation, as shown in Fig. 6.29. The available space of this new
casing has an outer diameter of 27.8 mm and inner diameter of 14.5 mm.
The initial design incorporated the following key features of an HF tag:
• A multi-turn loop to provide the required inductance.
Page 169
6.10 HF RFID Antenna Design
Part 1
Part 2
(a) (b)
Figure 6.28. Casing for livestock RFID tag. (a) Two parts of a cattle livestock tag so that it
can be pinned on the ears of cattle livestock. (b) The possible space for RFID tag
allocation on a cattle livestock tag: outer diameter of 28.3 mm and inner diameter of
12.0 mm.
Figure 6.29. Tag encapsulation casing from Leader Products. The dimension of spacing avail-
able: outer diameter of 27.7 mm and inner diameter of 14.5 mm.
• Allow as much space as possible in the middle of the coil to have a high induc-
tance value.
• An underpass and two via holes to complete the loop.
The antenna has 6 loops on one side of an FR4 substrate (1.6 mm thick) as shown in
Fig. 6.30. The formation of the multi-turn loop is achieved by connecting several almost
circular ring tracks together. In the simulation software, the circular shapes are formed
by polygons, which are visibly obvious in Fig. 6.30. Also, the connecting part is bent
and is not smooth.
Page 170
Chapter 6 RFID Tag Antenna Design
Figure 6.30. First version of a pig tag with bent tracks to connect all the circular tracks
together..
An alternative way to model the multi-turn loop antenna is to use a spiral structure. A
spiral multi-turn loop is shown in Fig. 6.31. There are no bend as compared to Fig. 6.30.
However, the simulation of a spiral model takes a comparatively longer time. Also,
similar to the circular shape, it is noticeable that the spiral track is formed by polygons
and the edges are not smooth.
Figure 6.31. Second version of a pig tag using a spiral track.
Hence, a spiral is not chosen as the final choice for the modelling of an HF pig ear tag.
Circular shapes are also avoided to prevent the polygonal representation. The solution
Page 171
6.10 HF RFID Antenna Design
is rather peculiar. Rather than using a polygonal circular shape, a higher order polygon
is defined. The choice of using a high order polygon can be seen in Fig. 6.32, which is
also the figure for the first prototype HF pig ear tag.
Figure 6.32. First prototype of a pig tag.
After the simulation, the antenna was sent for fabrication. A serious problem was
encountered during the fabrication. As shown in Fig. 6.33, both the gap between the
coil track and the width of the track itself of the HF antenna are very small. This often
results in coil track short-circuiting. Also, the uneven track increases the loss resistance
of the antenna. This will reduce the quality factor and reduce the performance of the
tag.
12
Figure 6.33. Problem of uneven track thickness when prototyping antenna with thin tracks.
(1) Thinner tracks, (2) Thicker tracks.
Page 172
Chapter 6 RFID Tag Antenna Design
After a satisfactory antenna was fabricated despite the problem mentioned above, a
simple test was carried out to check the resonance point of the fabricated antenna. A
small capacitor was soldered on to the feed point of the antenna to represent the input
capacitance of a HF RFID chip, as shown in Fig. 6.34.
Figure 6.34. First version prototype of a pig tag. No HF chip is attached. A capacitor of 100 pF
is used instead to represent the input reactance of a HF chip.
The added capacitor had a capacitance value of 100 pF, equal value to the HF chip input
capacitance of 100 pF (Section 6.6), to mimic an HF chip. The tag, HF antenna with the
added capacitor was then tested for its resonant frequency. It was then discovered that
the resonant frequency was 14.71 MHz.
This shows that with slight modification (slight increment in inductance) of the HF
tag antenna design shown in Fig. 6.34, the HF tag antenna will operate in the desired
frequency of 13.56 MHz. A suitable HF chip is the NXP-Phillips SL1 ICS31 01 ICODE1
Label IC [103] which has an input capacitance of 97 pF.
It should be noted that the test of tag resonant frequency by substituting an HF chip
with a capacitor is purely an early indication of whether the design approach is suitable
or not; this is achieved by measuring the resonance frequency. A test using an actual
HF chip is needed to confirm the functionality of the tag. As mentioned before in
Section 6.6, there are no available HF chip straps which resemble the UHF chip straps.
As discussed in Section 6.6, all the HF chips used in experiment are recycled chips from
existing HF tags. Hence, it is not feasible to use recycled HF chips at the beginning of
the design cycle. After simulation and fabrication, an easy way to test the design is to
Page 173
6.10 HF RFID Antenna Design
use a tag chip with Tag Talk First (TTF) compatibility (ISD72128, refer to Section 6.6).
A tag (same design with the one shown in Fig. 6.34 attached with a tunable capacitor
and a TTF HF chip is shown in Fig. 6.35.
Figure 6.35. Pig tag with TTF compatibility.
1
2
3
Figure 6.36. Testing a pig tag with TTF compatibility. (1) To a signal generator, (2) To a
reader antenna, (3) To a spectrum analyser.
Page 174
Chapter 6 RFID Tag Antenna Design
No RFID reader is required to test a TTF RFID tag. Once a TTF RFID tag is powered
by a CW wave at the correct frequency, it will start generating a signal, depending on
the induced voltage at its input terminal (See Section 6.6 for more details). Hence the
testing environment is as shown in Fig. 6.36, where a signal generator, a spectrum anal-
yser, an antenna and a three port directional coupler. The connection to the directional
coupler is as follows:
• Out: Connected to Signal Generator.
• In: Connector to the antenna of HF reader.
• Cpl: Connected to spectrum analyser to monitor the response from a tag.
This test shows a read range of approximately 0.2 m. The results prove that the tag is
working as an RFID tag. No fine-tuning was carried out to optimise the performance
of the tag as it was then decided to pick an HF chip with lower input capacitance. This
is due to the fact that the input capacitance value has great impact on the cost of an
HF chip as the chip will need bigger area to provide the necessary capacitance. Hence
it is most desirable to have a tag antenna design which can function on a lower input
capacitance HF chip, such as the 23.5 pF input capacitance RF-HDT-SJME-G1 Tag-it
HF-I transponder IC from Texas Instrument [102].
An HF antenna needs to provide sufficient inductance through its coil turns to resonate
with the tag chip capacitance at the frequency of operation, which is 13.56 MHz. In the
case of a size constraint and reduced HF chip input capacitance, an HF tag would not
have sufficient inductance, and extra capacitance would be added by making the HF
antenna double sided and introducing capacitance through the two overlapping plates.
To avoid the difficulties in fabrication (uneven tracks as shown in Fig. 6.33 caused
by the closely spaced tracks), the tracks are designed to be further apart as shown in
Fig. 6.37. Simulation was carried out using Ansoft HFSS to predict the performance of
the designed tag antenna.
The simulation results are as shown in Fig. 6.38. The resonant point of the HF an-
tenna alone is at approximately 40.2 MHz. The measurement results on a fabricated
prototype antenna show a resonant point at the frequency of 41.3 MHz.
Page 175
6.10 HF RFID Antenna Design
Figure 6.37. Pig tag with external capacitance added.
0.00 10.00 20.00 30.00 40.00 50.00-20000.00
-10000.00
0.00
10000.00
20000.00
30000.00
Freq [MHz]
Figure 6.38. Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of 1.6 mm.
(Red curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).
To have a high capacitance value, it is desirable for the tag to be as thin as possible to
reduce the distance between the overlapping coils and plates so that the capacitance
between the parallel plate in the middle and also the capacitance between the upper
and the lower rack of the HF coil are increased. The thickness of the FR4 was reduced
from 1.6 mm to 0.8 mm. The simulation results are as shown in Fig. 6.39. The simulated
resonant point is at approximately 29.2 MHz. The measurement results on a fabricated
prototype antenna show a resonant point at the frequency of 28.6 MHz.
Page 176
Chapter 6 RFID Tag Antenna Design
0.00 10.00 20.00 30.00 40.00 50.00-10000.00
0.00
10000.00
20000.00
Freq [MHz]
Figure 6.39. Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of 0.8 mm.
(Red curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).
Several simulations, antenna fabrications and tests were carried out to optimise the
performance of the tag. The final design of the HF antenna is shown in Fig. 6.41. The
thickness of the substrate has been reduced to 0.2 mm. Also, the coil tracks separa-
tion cannot be increased due to the size limitation. The chip that was used was the
Texas Instrument Tag-it HF-I Standard Transponder IC [102], which is based on the
ISO/IEC 15693 and ISO/IEC 18000-3 standards, and has an nominal input capacitance
of 23.5 pF.
12
3
Figure 6.40. Final version prototype of a pig tag. (1) Multi-turn coil to provide inductance, (2)
TI HF chip, (3) Tunable capacitance plate for fine tuning.
Page 177
6.10 HF RFID Antenna Design
The front and the back of the un-tuned final design is shown in Fig. 6.41.
(a) Front view (b) Back view
Figure 6.41. HF RFID Tag to be embedded in a livestock ear tag.
For testing purposes, the used ID ISC.LR2000 HF reader and ID ISC.ANT300 reader
antenna, both from FEIG Electronics were used. The width and length of the reader
antenna are both approximately 0.33 m. In HF operation, the maximum possible read
range is proportional to the size of the reader antenna [95]. As a rule of thumb in real
life deployment, the maximum read range is approximately equal to the dimension of
the reader antenna. A read range of 0.34 m is obtained in free space and of approx-
imately 0.32 m when attached to a human hand, which simulates the environment
where a tag is attached to the pig’s ear. To increase the read range, we can increase the
size of the reader antenna in a way presented in [95].
After encapsulation, a simple test was performed to certify that an encapsulated tag
was in fact made water proof. This simple test did not focus on any potential degrada-
tion of tag due to the presence of water. A tag was randomly selected and immersed in
water as shown in Fig. 6.43. The tag is constantly monitored over a long period as the
water container was positioned in the middle of an operating HF reader antenna (ID
ISC.ANT300/300-A) controlled by an HF reader (ID ISC.LR2000), both from FEIG Elec-
tronic. The immersed tag was functioning as usual which confirm the encapsulation
process was successful in preventing water damage to an RFID tag.
The deployment of HF RFID tags was carried out together with UHF RFID tags, where
the design of UHF RFID tags for pig tagging can be found in Section 6.11.4. The com-
plete report of RFID deployment in piggery can be found in Appendix G. Fig. 6.44
Page 178
Chapter 6 RFID Tag Antenna Design
(a) Before encapsulation (b) After encapsulation
Figure 6.42. RFID tags before and after encapsulation process. (a) HF (top) and UHF (bot-
tom) RFID tags embedded in the casing shown in Fig. 6.29. Discussion of UHF pig
tag can be found in Section 6.11.4. (b) An encapsulated RFID tag viewed from (from
left) front, side and back.
Figure 6.43. An encapsulated tag immersed in water to test for readability and water proof
capability. Tag remains functioning even after all the water has evaporated.
shows the overall RFID system setup in the pig pen in the piggery. Fig. 6.45 shows
the arrangement of HF and UHF RFID readers in protective casings in the piggery.
Fig. 6.46 shows the orientation of HF and UHF tags on pig ears, and Fig. 6.47 shows
the snapshot when pigs are feeding in an RFID enabled feeder.
Part of this section has been published in [118] and [119].
6.11 UHF RFID Antenna Design
This section presents the design of a UHF RFID antenna. A UHF RFID tag operates
within 860 MHz to 960 MHz around the world. A UHF RFID tag antenna is designed
Page 179
6.11 UHF RFID Antenna Design
Figure 6.44. Overall RFID system setup in piggery. Feeder on the left with UHF antennas and
feeder on the right with HF antennas.
Figure 6.45. HF and UHF RFID readers in piggery. Readers setup in piggery: HF (bottom) and
UHF (top) RFID readers in protective boxes.
Figure 6.46. Orientation of HF and UHF tags on pigs ears. A pig was tagged with one HF tag
(left in picture) and one UHF tag (right in picture).
to have a conjugate match with a UHF RFID chip to allow maximum power trans-
fer. Firstly, this section discusses the common commercial UHF RFID tags and their
common features. In the later part, various novel UHF antenna designs for various
applications are shown.
Page 180
Chapter 6 RFID Tag Antenna Design
Figure 6.47. Picture on pigs feeding in an RFID pig feeder. Pigs feeding at both the UHF (left
in picture) and HF (right in picture) feeders.
UHF tags, as discussed before, are very susceptible to the environment. Hence, the
design of UHF RFID tag antenna is very diverse, largely depending on its applica-
tion. There are numerous research papers on the design of UHF antennas for various
applications [120].
The design of UHF RFID antenna is not of the main research interest of the author. As
it can be seen later, most of the designs of UHF tag antenna come from joint research.
However, it is an important learning process as the design of a dual-frequency antenna
in Chapter 7 will require the design of UHF RFID tag antenna.
6.11.1 A Common UHF RFID Tag
A common commercial UHF RFID tag has very distinctive features. It is either a dipole,
or a small variation of the dipole, and has a simple matching network, most commonly
a matching network very similar to a T-Match is used to transform the capacitive na-
ture of a dipole into inductive to conjugately match to the input impedance of a UHF
RFID tag chip.
Also, common UHF RFID tags are very thin and flexible. They can readily be applied to
a pallet , or carton as used in retail supply chains. They can have a read range of more
than 10 m, in some regulatory regions, using the air interface specified by EPCglobal
C1G2 [35].
Page 181
6.11 UHF RFID Antenna Design
6.11.2 Simple Planar Dipole
This section presents the fundamental principles of operation of a simple planar dipole.
A dipole or electric dipole is the dual of a loop antenna or magnetic dipole. As sug-
gested by the name, an electric dipole antenna is sensitive to electric field while a mag-
netic dipole antenna (loop antenna) is sensitive to magnetic field. As a common name,
an electric dipole is often known as a dipole while a magnetic dipole is known as a loop
antenna. A dipole and a loop antenna have very similar radiation patterns (doughnut
shape) when correctly aligned.
A half wave length dipole is theoretically purely resistive (zero reactance). However in
actual implementation, a half wave dipole has to be shortened by about 4-10 percent to
present a purely resistive impedance to the feed point. From [85], the input impedance
of a half wave length dipole, Zin, is equal to 73 + j42.5 Ω with a directivity of 1.643.
In theory, as mentioned above, a half wave dipole is purely resistive. A dipole shorter
than a half wave dipole is always capacitive. Increasing the size of dipole past the
purely resistive point will push the impedance into the inductive region. An inductive
dipole without any matching network is generally significantly longer than a half wave
length.
The UHF band for RFID operation around the world is from 860 MHz to 960 MHz. A
half wave length dipole for this frequency band is between 0.3125 m and 0.3488 m. A
typical UHF RFID tag is normally around 0.1 m in length, and hence, a dipole without
any matching network is not feasible to be used as a UHF RFID tag antenna.
Some of the popular matching networks for a dipole are known as the Omega and
Gamma matching networks. However, an Omega matching network can only match
loads less than 50 Ω resistive by increasing the resistance. An Omega matching cannot
step or transform antenna feed resistance downwards when using capacitors. Also,
both Omega and Gamma matching networks cannot match capacitive antenna loads.
A variation of the dipole is the folded dipole. The directional characteristics of a folded
dipole are the same as those of a simple dipole. However, the reactance of a folded
dipole varies much more slowly as the frequency is varied from resonance. Because of
Page 182
Chapter 6 RFID Tag Antenna Design
this the folded dipole can be used over a much wider frequency range than is possible
with a simple dipole. For RFID applications, the dipole, folded or not, is always planar.
6.11.3 A UHF Antenna for Sheep
This section presents a UHF antenna for sheep tagging. This section is an extract from
a co-authored paper “A small passive UHF RFID tag for livestock identification” [121].
Similar to pig tagging as discussed in Section 6.10.4, animals such as cattle and sheep
are tagged for purposes such as disease control, breeding management and also stock
management. Also similar to the pig tagging, where a passive HF RFID tag is attached
on a pig’s ear, sheep tagging uses a small passive UHF RFID tag to be attached on a
sheep’s ear.
The current RFID technology for livestock tagging, including both pig tagging and
sheep tagging, operate in the low frequency (LF) region [116]. However, as discussed
in Section 6.10.4, LF tags can only be read at close range and may not perform well
when multiple tags are simultaneously present in the interrogation field.
It is believed that the HF RFID tag designed for pig tagging as presented in Sec-
tion 6.10.4 can be used in sheep tagging as well. However, the focus of the research
on sheep tagging is on UHF operation. As compared to HF tags, UHF tags not only
give better read range, but also support higher data rates.
The design of the UHF RFID sheep tag is based on a loop antenna, contrary to the
common dipole built UHF RFID tag. The main reason is that a loop will provide in-
ductance useful in the matching to the capacitive input impedance of a tag chip. Also,
a loop antenna does not have sharp edges and can be fitted on the sheep’s ear relatively
easily.
This sheep tag is designed to operate in the frequency of 915 MHz. (Note: This fre-
quency is not normally allowed in Australia. The test was carried out at that frequency
but in a controlled environment.). The UHF chip used is the Alien C1G1 while the
reader used is the Alien UHF RFID reader.
Page 183
6.11 UHF RFID Antenna Design
Figure 6.48. Sheep ear tag. (a) Front view; (b) Rear view of the sheep ear tag
The front and rear view of the final product is shown in Fig. 6.48. The matching net-
work is provided by the copper strips on the rear side, providing series and shunt
capacitance to the circuit. These copper strips can be trimmed to fine tune the circuit,
during development, so that the antenna will resonate at the desired frequency, which
is 915 MHz. This antenna is designed to operate in the USA, but with a simple tuning
(trimming of copper strips on the rear side of the antenna), it can operate in Australia
conforming to the local regulations.
The target impedance of the designed antenna is 6.5371 + j133.44 Ω, as compared to
the chip impedance of 17.36 + j149.2 Ω at 915 MHz. The maximum measured read
range was about 1.42 m.
However, as predicted by the radiation pattern, when the tag was aligned with its
axis parallel with that of the interrogator antenna, the reading distance was only a few
centimetres. A simple example of the possible the deployment of this tag for sheep
Page 184
Chapter 6 RFID Tag Antenna Design
tagging is to locate the reader antenna on top of a door arch where the tagged animals
pass through, with the direction of the reader antenna facing down.
6.11.4 A UHF Antenna for Pigs
This section is based on a published paper “Investigation on the deployment of HF and
UHF RFID tag in livestock identification” [118]. The background discussion of RFID
tags for pigs is presented in Section 6.10.4.
Two different UHF tag designs have been considered in our study. Both UHF tags have
been designed for operation at 923 MHz in Australia. The tag chips used for both tags
are from Alien Technology [122], and have impedance of approximately 20 − j141 Ω at
923 MHz.
25
mm
2 mm
copperstrip
substrate
chip
FRONT
13
mm
1 mmseriescapacitor
shuntcapacitor
BACK
x
y
(a) Tag 1(b) Tag 1: radiation
pattern
28m
m
2 mm
13
mm
0.5 mm
chip
x
y
(c) Tag 2(d) Tag 2: radiation
pattern
Figure 6.49. Two UHF pig tags.
The first UHF tag design consists of a circular loop antenna with a two element match-
ing network to match the impedances of the tag antenna and tag chip. More detailed
design steps of this tag can be found in [121]. However, the tag has been made slightly
larger in diameter compared to that presented in [121] to suit the application studied in
Page 185
6.11 UHF RFID Antenna Design
Table 6.5. Read range for UHF tags.
UHF Tag 1 UHF Tag 2
Tag in free space (m) 1.00 0.48
Tag attached to hand (m) 0.20 0.27
this research. The structure and dimensions of this tag is shown in Fig. 6.49(a). The tag
is made using a thin FR4 board with substrate thickness h = 0.36 mm and relative di-
electric permittivity εr = 4.4. This tag design utilises both sides of the FR4 board. From
Fig. 6.49(a), the front view is the circular loop antenna and the back view is the imple-
mentation of the matching network that consists of a series and a shunt capacitor. The
tag antenna structure is simulated using Ansoft HFSS and the simulated impedance is
10 + j148 Ω, which is the conjugate of the tag chip impedance. The radiation pattern
of the antenna is as shown in Fig. 6.49(b), with maximum directivity of 1.65 dB.
The second UHF tag design considered consists of a curved electric dipole antenna
with an inductance track across the dipole for impedance matching purposes. The
structure and dimensions of this tag is shown in Fig 6.49(c). This tag is made using an
FR4 board with substrate thickness h = 1.6 mm and relative dielectric permittivity εr =
4.4. This tag design is single-sided. The tag antenna structure is simulated using Ansoft
HFSS and the simulated impedance is 1 + j152 Ω. As can be seen, the impedance of
the tag antenna structure is not exactly the conjugate of the tag chip impedance and
hence, the tag antenna and chip impedances are not perfectly matched. However,
the inductance track across the dipole does provide sufficient inductance to tune with
the capacitance of the tag chip. The radiation pattern of the antenna is as shown in
Fig. 6.49(d), with maximum directivity of 1.83 dB. As can be seen, although the general
shape of the radiation pattern of the first and second UHF tags are almost the same,
the maximum directivity of these two tags occur at different directions with respect to
the plane of the tags. Both UHF tags are tested and the read range results are shown in
Table 6.5.
The complete report of RFID deployment in piggery can be found as Appendix G.
Page 186
Chapter 6 RFID Tag Antenna Design
6.11.5 A UHF Antenna for Beer Kegs
This section presents a UHF antenna for a Beer Keg. This section is an extract from a co-
authored paper “A Small Passive UHF RFID Tag for Metallic Item Identification” [123].
The tag presented here is not only for the deployment of RFID in beer keg tagging but
the main purpose is to illustrate how an RFID tag can be designed to be placed on a
metallic surface. The idea of this tag design is to fully exploit the fact that magnetic
field is doubled at a metallic surface. The loop antenna is oriented in such a way as
to couple to the magnetic field at the metallic surface. The final design is as shown in
Fig. 6.50.
Tag antenna
Tag
chip
Width (W )rec Length ( )Lrec
Height
(H )rec
x y
z
Figure 6.50. Structure of the RFID tag with a rectangular loop antenna.
The simulated antenna impedance is shown in Table 6.6. The read range of 0.83 m
is obtained when a reader is operating at reduced power. The reason for testing at re-
duced operating power, is that this application is likely to be used with hand held RFID
readers, and to conserve battery power, the RF output power is limited to 500 mW. It
is expected the tag can be read up to 1.3 m when a reader is operating at maximum
allowable power.
Table 6.6. Read range for UHF tag for beer keg.
Antenna Impedance Peak Directivity
Free Space 0.26 + j91 Ω 1.5 dB
Above Metal 0.53 + j91 Ω 6.5 dB
Page 187
6.12 Conclusion
6.11.6 A UHF Antenna for New Wine Closures
This section describes the design of a UHF antenna for a wine closure. The wine closure
of interest is the same as presented in Section 6.10.3 and illustrated in Fig 6.26.
Figure 6.51. RFID tag for wine closure.
The functionality of the UHF tag for the wine closure is also the same with the the
HF tag for wine closure as discussed in Section 6.10.3. The final design is as shown in
Fig. 6.51 and the details and result of actual testing is presented in Table 6.7.
Table 6.7. Performance of UHF RFID tag for the Zork wine closure.
Material Copper strip
Dimension (mm) 6.0 (h) 7.0 (r)
Chip Alien Class 1 Generation 1
Reader Alien ALR-9780
Read Range (mm) ≈ 300
6.12 Conclusion
This chapter presents the fundamentals of HF and UHF tag designing, novel HF and
UHF RFID tag antenna designs for new applications, antenna modelling, simulation
Page 188
Chapter 6 RFID Tag Antenna Design
and prototyping, and tag fabrication process. This provides a strong foundation to
advance into the design of dual-frequency antenna for RFID application, which is es-
sentially a single antenna operating as both a HF antenna and a UHF antenna.
Page 189
Page 190
Chapter 7
Dual-Frequency RFIDAntenna
WITH good read range and fast data rates, UHF Radio
Frequency Identification (RFID) is being deployed in the
supply chain to uniquely identify each item. However,
UHF transmission power is easily absorbed by ionised liquids, such as wa-
ter. Hence, normally, in an environment with lots of liquid products, an
HF RFID system is chosen instead. This chapter investigates the feasibility
of designing a dual-frequency (HF and UHF) RFID tag antenna, with fre-
quency ratio of up to 70, to embrace the benefits offered by both the UHF
and HF RFID systems. Furthermore, this chapter shows that with care-
ful design, a single feed dual-frequency RFID tag antenna can be achieved.
A miniaturised version of a dual-frequency RFID tag antenna is also de-
signed. Prototypes of the dual-frequency RFID tag antenna are presented
in this chapter, along with validating simulations and measurement results.
Page 191
7.1 Introduction
7.1 Introduction
As discussed in Chapter 2, Radio Frequency Identification (RFID) is a technique used
to identify objects by means of electromagnetic waves. An electronic code responding
label, also known as a tag, consists of an antenna and an integrated circuit. Upon
receiving any valid interrogating signal from any interrogating source, such as an RFID
reader, the tag will respond according to its designed protocol.
RFID has been recently applied into supply chain tracking systems, revolutionising
the conventional ways of how an object is tracked along the supply chain and how an
inventory system works. Not only does RFID bring benefits in saving costs, it also has
the potential of increasing security levels along the supply chain, especially in genuine
goods authentication. To reduce the cost of RFID deployment, passive tags are used.
There are 4 common bands for RFID applications; these are the LF band (less than
135 kHz), the HF band (13.56 MHz), the UHF band (860 - 960 MHz) and the so-called
microwave (2.45 GHz). LF is not suitable for supply chain deployment, as to uniquely
identify every object in the supply chain, each tag has to bear a unique EPC number,
which must be at least 64 bits in length [4]. A low frequency operation, such as in
the LF band, would suffer very slow reading in a heavily populated tag environment.
Also, LF requires large antenna components and hence is difficult to implement and
is susceptible to electrical noise, which HF can handle [20]. UHF and microwave tags
can offer comparatively very fast reading, but their performance will suffer more than
the other bands described above in the presence of liquid or metal [19]. It is very dif-
ficult to conclude whether HF or UHF is better for RFID application in supply chains.
This is because in any supply chain, there are many different scenarios, and in some
HF is better and in the others, UHF outperforms HF operation. Table 7.1 summarises
comparison between HF and UHF as presented in [15, 19, 24].
To embrace both the advantages contributed by HF and UHF, this chapter investigates
the feasibility of designing a dual-frequency antenna for an RFID tag. The next sec-
tion reviews current technology on dual-frequency antennas, and their shortcomings
in designing an RFID HF and UHF dual-frequency antenna. Design aims ensuring
Page 192
Chapter 7 Dual-Frequency RFID Antenna
Table 7.1. HF versus UHF in RFID operation.
HF UHF
Read Range ≈ 1 m Up to 10 m
Cost (in large volume) Medium Low
Read Rate Low High
Metallic Area Bad Bad
Liquid Surrounding Better Worse
that the antenna is feasible for real life application are presented in Section 7.3. Sec-
tion 7.4 describes the design process in detail, including the problems encountered
and the solution of those problems using simulation results. Section 7.5 presents the
actual measurement results for the designed dual-frequency antenna, together with
some discussions and comments.
Section 7.6 begins the redesign of the novel dual-frequency antenna as presented in
Section 7.4, with the emphasis on miniaturisation. Also included in this section, are
the characteristics of the novel dual-frequency antenna (Section 7.4) retained for the
new miniaturised antenna design. The new or improved characteristics are discussed
in Section 7.7. Section 7.8 explains the test procedures used to fine-tune and verify
the functionality of the new miniaturised dual-frequency antenna while Section 7.9
shows the final design with its dimensions. This chapter is based on two published
paper [77, 124].
7.2 Current Dual-Frequency Antenna Design
The intended dual-frequency RFID antenna is planned to function at 13.56 MHz (HF)
and also within the UHF band for RFID operation (860 - 960 MHz). If using the high-
est frequency in the UHF band, which is 960 MHz, the frequency ratio for this dual-
frequency antenna is slightly more than 70.
Page 193
7.3 Design Aims
Wong has undertaken extensive research on dual-frequency microstrip antennas, with
the frequency ratio ranging from 1.8 - 4.9 [125]. However, the frequency ratio of this
technique is too low to be applied to the intended RFID dual-frequency antenna. A
common aperture, dual-feed dual-frequency antenna for 900 MHz and 60 GHz was
presented by Menzel [126]. The frequency ratio achieved was approximately 70. How-
ever, the antenna had dual feed points.
One of the antennas which has a high frequency ratio and used in RFID is the one
produced by IPICO [127]. Under the IPICO dual-frequency system, the RFID reader
transmits a signal using the LF spectrum (125 kHz to 135 kHz) in order to power the
tag and the tag uses the HF spectrum (3 to 30 MHz) to transmit its signal back to the
reader. The tag produced by IPICO is using LF and HF, while the aim of this chapter is
to combine both HF and UHF.
7.3 Design Aims
The aim is to design, simulate and fabricate a dual-frequency RFID antenna, which can
demonstrate the following characteristics:
1. Antenna impedance is equal to the complement of the input impedance of the RFID chip
at UHF operation.
There is no available dual-frequency RFID chip to be used as the guidelines for
the design of dual-frequency RFID tag antenna. However, in a hypothetical
way, it is predicted that a dual-frequency RFID tag chip will have a similar in-
put impedance at UHF as the input impedance of a commercially available UHF
RFID tag chip (such as those presented in Section 6.6).
With a dual-frequency antenna having the impedance complement to a chosen
RFID chip at UHF, maximum power can be transferred from the antenna to the
RFID chip. The RFID chip used in the fabrication process has an approximate in-
put impedance of 17− 150j Ω at RFID UHF band. Hence the final antenna design
must have an input impedance of approximately 17 + 150j Ω. In actual fact, the
chip impedance will change within the RFID UHF band, but in this chapter the
Page 194
Chapter 7 Dual-Frequency RFID Antenna
UHF antenna is designed to be matched at the highest RFID operating frequency
(960 MHz). It is believed that if an antenna can work at the highest frequency
(having the highest frequency ratio), the antenna can be tuned in a reasonably
easy way to operate at the lowest frequency (860 MHz).
2. A resonance point at HF.
Similar to point 1, where is no available dual-frequency RFID chip to be used as
the guidelines for the design of dual-frequency RFID tag antenna at UHF, there is
no guideline for the design of HF functionality. Again, in a hypothetical way, it is
predicted that a dual-frequency RFID tag chip will have a very small (negligible
level) of input capacitance at HF. This is not the case for a conventional HF chip.
From the HF chip examples from Section 6.6, the HF chip can have an input
capacitance from 23.5 pF up to 97 pF.
Hence, the designed dual-frequency antenna must have sufficient inductance
and capacitance to have a resonant point at 13.56 MHz. Also, in order to have
a high quality factor, the resistive losses must be kept as low as possible.
3. A single feed antenna.
The final antenna design is planned to be attached to commercially available
RFID chips. Hence a single feed antenna is desired to avoid any modification
required to the current RFID chips’ structure, where only a single feed point ex-
ists. Although there is no existing dual-frequency RFID tag chip available, it is
envisaged that a single feed dual-frequency RFID tag chip will have a lower cost
than a multi-feed dual-frequency RFID tag chip.
4. Reasonable antenna size and cost.
The material used for fabrication is FR4, with relative permittivity, εr, equal to 4.4.
The antenna area should not exceed 14400 (mm)2. However, it is not the focus at
the initial stage to minimise the size of the antenna. Hence, the first novel design
may not be of the smallest possible size. The design of a miniaturised version of
a dual-frequency RFID tag antenna can be found in Section 7.6 onwards.
Page 195
7.4 Dual-Frequency Antenna Design
7.4 Dual-Frequency Antenna Design
Through brainstorming, three main methodologies to design a dual-frequency antenna
were identified:
1. Merging an HF antenna and a UHF antenna.
2. Making an HF antenna operate in UHF.
3. Making a UHF antenna operate in HF.
This chapter will focus on the first methodology, while Chapter 8 will present the re-
search results from the other methodologies.
The design philosophy used to design a dual-frequency antenna (HF and UHF), is to
design an HF and a UHF antenna separately as in Section 7.4.1, and merge these two
antennas together with a single feed. However, since these two antennas are joined
together in parallel, the UHF antenna will affect the HF operation of the HF antenna,
and vice versa. A few experiments were carried out to verify the merging idea and are
presented in Section 7.4.2 and Section 7.4.3.
7.4.1 Independent HF and UHF Antenna Design
For operation at HF, a multi-turn planar loop antenna is designed. The calculation of
the inductance of a multi-turn planar loop antenna is done using [113], which includes
the computation of the positive and negative mutual inductances. From f0 = 12×π
√LC
,
with an intended resonant frequency of 13.56 MHz, we know that a large inductance
in required. The problem is at the later stage, when this loop antenna is merged with
the UHF dipole, the resonant frequency will change. Hence, the number of turns of
the multi-turn loop is not fixed at this stage, and will only be finalised during the fine-
tuning stage after the merging of the HF and the UHF antennas.
An example of a generic multi-turn planar loop antenna is shown in Fig. 7.1. For oper-
ation at UHF, a dipole is used. An example of a generic half wave-length (of 960 MHz)
Page 196
Chapter 7 Dual-Frequency RFID Antenna
dipole is shown in Fig. 7.2. Note that the feed points of the antennas are represented by
small rectangular boxes in those figures. This is the case for all the antennas presented
in this thesis.
Figure 7.1. A generic HF coil antenna. The multi-turn loop provides the required inductance.
Figure 7.2. A generic UHF dipole antenna.
Page 197
7.4 Dual-Frequency Antenna Design
7.4.2 Quick Feasibility Test
To prove the idea of merging an HF antenna together with a UHF antenna, an HF an-
tenna is merged with a UHF antenna in HFSS to observe the feasibility and to form the
foundation of building a dual-frequency RFID tag antenna which will operate accord-
ing to specifications.
Generic models from Section 7.4.1 are used. A simple four-loop planar coil antenna is
used as the HF antenna and a simple dipole is used as the UHF antenna. Two via holes
and a underpass are required for the connectivity of the simple four-loop planar coil
antenna. The model of this dual-frequency antenna is shown in Fig. 7.3.
The width of the track is arbitrarily chosen as 3 mm with the gaps between the HF
planar coil loops chosen as 1 mm.
The antennas are interconnected by a transmission line as shown in Fig. 7.3. The length
of the transmission line is 80 mm, with the HF antenna connected to one side, and the
UHF antenna on the other side. This length includes the 3 mm track of both the HF
and the UHF antennas at the junctions where these two antennas are connected to the
transmission line. The length was chosen to be approximately λ4 , which is 78 mm (for
a frequency of 960 MHz). Although the permittivity of the substrate will affect this
length, it is not considered at this stage, but will be discussed in detail in later stages.
The length of dipole including the gap is 132 mm. The choice of length is based on
the fact that a practical half-wave dipole is shorter than its theoretical half-wavelength
which is 156 mm (at a frequency of 960 MHz). Again, the effect of the substrate is not
taken into account at this stage.
This initial simulation of a dual-frequency antenna with the proposed method of merg-
ing two different antennas together showed promising results. Fig. 7.4 shows a reso-
nant point at HF (32.5 MHz). Although the obtained resonant point is not exactly at
the intended frequency of 13.56 MHz, the addition of inductance or/and capacitance
will provide the correct resonant frequency. Fig. 7.5 shows that the resistance is higher
than the target value while the reactance is capacitive (the target reactance should be
inductive). Using a matching network, the impedance characteristics at UHF can be
Page 198
Chapter 7 Dual-Frequency RFID Antenna
Figure 7.3. First version of a dual-frequency antenna. A generic HF coil antenna as shown in
Fig. 7.1 and a generic dipole as shown in Fig. 7.2 are connected together with a single
feed point.
0.00 10.00 20.00 30.00 40.00 50.00-10000.00
-5000.00
0.00
5000.00
10000.00
15000.00
Freq [MHz]
Figure 7.4. Impedance of the first version of dual-frequency antenna at HF. (Red curve is
the resistance curve and blue curve is the reactance curve, Y-axis in Ω). The fre-
quency response at HF of the antenna shown in Fig. 7.3. The resonant point occurs at
32.5 MHz.
easily matched to the tag chip impedance for the intended UHF frequency band as
shown in Section 7.4.4.
This quick feasibility test shows that the merging of two independently designed an-
tennas using a transmission line and a single feed is possible. Fine-tuning is required
Page 199
7.4 Dual-Frequency Antenna Design
860.00 880.00 900.00 920.00 940.00 960.00-100.00
-50.00
0.00
50.00
100.00
150.00
Freq [MHz]
Figure 7.5. Impedance of the first version of dual-frequency antenna at UHF. (Red curve is
the resistance curve and blue curve is the reactance curve, Y-axis in Ω). The frequency
response at UHF of the antenna shown in Fig. 7.3. The antenna is capacitive across
the UHF band of interest.
for this dual-frequency antenna to operate in the desired frequency bands. A simple
test was carried out (Section 7.4.3) to check the tunability of this antenna.
7.4.3 Tunability Test
The first version of the dual-frequency antenna (Fig. 7.3) was then modified slightly to
test its tunability and is named second version of the dual-frequency antenna (Fig. 7.6)
for easy referencing. This quick tunability test is to demonstrate whether the character-
istics of the dual-frequency antenna can be changed in a predicable way when simple
modifications are applied.
In this test, the dipole is bent to reduce of the overall size of the dual-frequency an-
tenna. Also, the length of the transmission line linking the HF and UHF antennas is
varied (10, 20, 30, 40, and 80 mm), to observe the changes in impedance. An example of
the simulation model used is shown in Fig. 7.6 where the dipole is bent and the length
of the transmission line is 10 mm.
Page 200
Chapter 7 Dual-Frequency RFID Antenna
Figure 7.6. Second version of dual-frequency antenna. The overall structure of the first version
of dual-frequency antenna is redesigned, with the shortening of the transmission linking
the HF and the UHF antenna, and the bending of the UHF dipole to reduce space.
An example of the graph showing the results obtained through simulation are shown
in Fig. 7.7, which corresponds to the simulation of the model in Fig. 7.6. This impedance
curves looks similar to the impedance curves of the first version dual-frequency an-
tenna (Fig. 7.5). The slight difference is the the reactance curve which has been shifted
up, and in parts of the frequency band, it has become inductive where it was previously
capacitive.
• Prediction of Impedance
Since the HF and the UHF antennas are connected in parallel, it might be possible
to predict the overall impedance if the individual impedance of both the HF and
UHF antennas are known. This test is to examine the coupling effect between the
HF and UHF antennas when placed near to each other.
The idea is to assume the antenna impedances are lumped and a simple parallel
circuit can be formed and the combined impedance, Zcal, can then be calculated
using Zcal = Zdipole||Zcoil, where Zdipole and Zcoil are the simulated impedances
of the individual UHF antenna and HF antenna respectively. If there is negli-
gible coupling effect between these two antennas, Zcal should be equivalent to
the simulated impedance for the dual-frequency antenna, Zsimu. However, the
Page 201
7.4 Dual-Frequency Antenna Design
0.85 0.90 0.95 1.00 1.05-400.00
-200.00
0.00
200.00
400.00
600.00
Freq [GHz]
Figure 7.7. Impedance of the second version of dual-frequency antenna at UHF. (Red curve is
the resistance curve and blue curve is the reactance curve, Y-axis in Ω). The frequency
response at UHF of the antenna shown in Fig. 7.6. The reactance curve shows that
it is possible to push the reactance curve into the inductive zone with minimal design
modification.
requirement of a transmission line to link the HF and the UHF antennas together
will affect the results. This will be taken into consideration when making a con-
clusion on this simple test.
Table 7.2 shows the calculated combined impedance of a dual-frequency antenna
when given individual impedances of an HF antenna and a UHF antenna while
Table 7.3 compares the calculated combined impedance of a dual-frequency an-
tenna with the simulated impedance of the same dual-frequency antenna. Fig. 7.8
shows clearly the accuracy of the calculated impedance value as compared to the
simulated impedance based on the values shown in Table 7.3.
It can be seen from Fig. 7.8 that the calculated impedances are close to the sim-
ulated impedances, following a similar pattern. The inaccuracy occurs mostly
around the resonant zone (between 0.85 - 0.95 GHz), and can be improved if
more frequency points are considered.
Page 202
Chapter 7 Dual-Frequency RFID Antenna
Table 7.2. Calculation of combined impedance based on individual impedances. Zdipole and
Zcoil are the simulated impedances of the individual dipole and coil antenna used to form
the dual-frequency antenna shown in Fig. 7.6. Zcal is obtained by summing in parallel
Zdipole and Zcoil.
Freq. Re(Zdipole) Im(Zdipole) Re(Zcoil) Im(Zcoil) Re(Ztotal) Im(Ztotal)
(GHz) (Ω) (Ω) (Ω) (Ω) (Ω) (Ω)
0.850 162.8325 122.4763 44.5804 -79.2678 71.2344 50.7455
0.875 205.5780 141.7485 106.0059 -19.0146 80.2015 -4.0880
0.900 260.7286 153.8072 114.7911 -130.8109 130.1086 51.7748
0.925 329.0677 151.2801 55.9386 -97.2257 75.9360 71.7810
0.950 406.5400 123.7083 44.1677 -60.3592 49.4806 49.2760
0.975 479.8166 61.5165 47.0725 -24.9616 44.3721 20.3141
1.000 526.5993 -33.9802 70.6423 -6.4963 62.2915 5.5255
1.025 528.5103 -142.8569 67.1498 10.2945 60.5988 -6.5156
1.050 487.4053 -236.7669 83.2981 50.2712 80.6469 -34.7300
Nonetheless, this experiment confirms that the total impedance of a single feed
dual-frequency antenna is dependent on both the impedance of the combined an-
tennas following simple parallel circuit theory. The small difference between Zcal
and Zsimu is caused by the addition of the transmission line and slight coupling
between the HF and the UHF antennas.
• Test of Transmission Line Length
As mentioned before, a transmission line is used to connect the HF and the UHF
antenna together with a single feed point to form a dual-frequency antenna. This
second quick feasibility test examines how the length of the transmission line can
affect the input impedance of the dual-frequency antenna. This transmission line
of interest is shown in Fig. 7.9.
The single feed point is to be located at one of the original feed points as shown
in Fig. 7.9(a). The decision to choose either point A or point B to be the feed point
is based on the following arguments:
Page 203
7.4 Dual-Frequency Antenna Design
Table 7.3. Comparison between calculated impedances and simulated impedances. Zcal is
the computed impedance of the dual-frequency antenna shown in Fig. 7.6, by summing
in parallel the impedance of the individual dipole and coil antenna used to form the dual-
frequency antenna shown in Fig. 7.6. Zsimu is simulated impedance of the dual-frequency
antenna shown in Fig. 7.6.
Freq. Re(Zcal) Im(Zcal) Re(Zsimu) Im(Zsimu)
(GHz) (Ω) (Ω) (Ω) (Ω)
0.850 71.2344 -50.7455 181.7573 39.9317
0.875 80.2015 4.0880 178.9783 121.3906
0.900 130.1086 -51.7748 345.1479 -215.3287
0.925 75.9360 -71.7810 93.2314 -137.3622
0.950 49.4806 -49.2760 60.5289 -67.6305
0.975 44.3721 -20.3141 56.5010 -22.3317
1.000 62.2915 -5.5255 74.0738 10.1240
1.025 60.5988 6.5156 65.7329 17.8289
1.050 80.6469 34.7300 80.8109 75.9376
– Consider the case where point A is chosen as the single feed point. If the
UHF antenna acts as an open circuit at HF, the transmission line needs to be
as short as possible, so that the open circuited UHF antenna would not in-
terfere with the HF antenna at HF. If the UHF antenna acts as a short circuit
at HF, the length of the transmission line needs to be λ4 (HF) (approximately
5.5 m), which is physically too large to be implemented and attached on
most of the item. At UHF, the HF antenna will interfere with the UHF an-
tenna, and depending on the length of the transmission line, it could be very
difficult to fine-tune the antenna to obtain the matching impedance.
– Consider the case where point B is chosen as the single feed point. If the HF
antenna appears to be an open circuit at UHF, the transmission line must
be as short as possible so that it will still looks as an open circuit even with
the addition of the transmission line. If the HF antenna appears to be an
Page 204
Chapter 7 Dual-Frequency RFID Antenna
-300
-200
-100
0
100
200
300
400
0.85 0.9 0.95 1 1.05
Freq (GHz)
Imp
eda
nce
(O
hm
)
Re(Simu)
Re(Cal)
Im(Simu)
Im(Cal)
Figure 7.8. Graph showing the calculated impedance and simulated impedance of the second
version of dual-frequency antenna. The calculated impedance, using the summation
of parallel impedances, has similar pattern with the simulated impedance. The difference
in results is caused by the coupling effect between the HF coil antenna and the UHF
dipole.
short circuit at UHF, the transmission line needs to be λ4 (UHF) (approxi-
mately 78 mm at 960 MHz and approximately 87 mm at 860 MHz), which
is realisable as an RFID tag antenna. At HF, the UHF antenna can be used
to provide additional capacitance to the HF antenna in obtaining a resonant
point at 13.56 MHz.
l
Coil Antenna
Dipole Antenna
A B
(a)
Bl
Merging of Antenna(Feed Point at B)
(b)
Figure 7.9. Location of feed point. (a) A transmission line is added between the two antennas.
(b) Feed point located at B.
Page 205
7.4 Dual-Frequency Antenna Design
From these arguments, it is more sensible to design a dual-frequency antenna
with the feed point located at B (Fig. 7.9(b)). Hence feed point B is chosen, with
the length of the transmission line, l, equals λ4 (UHF). The loop antenna (for HF)
is to be designed to have low impedance at UHF, to be transformed into high
impedance using the λ4 (UHF) transmission line, so it will not interfere with the
UHF dipole. The UHF dipole should be capacitive at HF and, as mentioned be-
fore, will contribute the capacitance to the HF antenna.
As discussed when determining the feed point location, the length of the trans-
mission line is always fixed to λ4 , in order to transform a low impedance to high
impedance, and vice versa. However, even though it is believed at this stage that
the HF antenna will have a low impedance at UHF, it might not be as low as de-
sired, so that when transformed it will not become very high. In this simple test
of transmission line length, it is hoped that a small variation in the length will
result in a small and predictable change in impedance so that in any case where
fine-tuning is needed, the length of transmission line can be considered.
The simulated impedances of the dual-frequency antenna with different trans-
mission line length are plotted in a Smith Chart shown as Fig. 7.10. Simulations
were carried at various discrete frequency points and hence the plots are jagged
and not smooth. However, Fig. 7.10 provides sufficient insight to shows that
there is a pattern in the changes of impedance when the length is varied.
Both the simple examinations on the tunability of the to be designed dual-frequency
antenna returned positive results. Hence, the focus will be diverted to the actual design
of this planned dual-frequency antenna, with single common feed point located at
point B (Fig. 7.9(b)).
7.4.4 Redesign of the UHF Dipole
In the previous section, two important facts have been shown: (1) A single-feed dual-
frequency antenna can be obtained by merging two independent antennas, and (2) an
adjustable transmission line can be used to control the overall impedance.
Page 206
Chapter 7 Dual-Frequency RFID Antenna
10mm
20mm
30mm
40mm
80mm
Figure 7.10. Smith Chart showing the simulated impedance of the dual-frequency antenna
with different transmission line length. As can be seen, as the transmission line
length changes (from 10 mm to 80 mm, represented by different colours as shown
in the top left corner of the Smith Chart), the impedance will change accordingly,
following a vague but definite pattern.
However, as can be seen from Fig. 7.5 and Fig. 7.7, the dipole is not matched to the RFID
chip impedance. It is not easy, and even may be impossible, to vary just the dimension
of the dipole to obtain the needed input impedance to have a perfect match.
Hence, the UHF dipole is redesigned so that it operates in the frequency band as in-
tended in a matching condition. The frequency band for RFID around the world is
within 860 - 960 MHz.
The first design is shown in Fig. 7.11. It is just a normal dipole with a matching network
added. The matching network chosen is a modified version of the T-match as discussed
in [85].
The designed antenna is tuned carefully to have a good match with a UHF chip for
a large bandwidth. The design is shown in Fig. 7.12, and the impedance within the
Page 207
7.4 Dual-Frequency Antenna Design
Figure 7.11. A UHF dipole with a matching network. The matching network is a modified
T-match network.
UHF RFID band is shown in Fig. 7.13. However, the real part is slightly higher than
expected. Also, the dimension of the dipole is too long in the x-axis direction.
Figure 7.12. A UHF dipole with tuned matching network. The matching network (A modified
T-match) can be fine-tuned to operate in the band of interest. In our case, it is tuned
to conjugately match to a UHF RFID tag chip at 960 MHz.
Hence, the two sides of the dipole are bent as shown in Fig. 7.14. When comparing with
Fig. 7.12, the length at the x-axis is reduced from 168 mm to 78 mm, at the expense of the
overall length in the y-axis. However, this will not be a problem, as the transmission
line has to be positioned along the y-axis as well. The impedance of this new design is
Page 208
Chapter 7 Dual-Frequency RFID Antenna
0.85 0.90 0.95 1.00 1.05 0.00
50.00
100.00
150.00
200.00
250.00
Freq [GHz]
Figure 7.13. Impedance of the UHF dipole with tuned matching network. (Red curve is the
resistance curve and blue curve is the reactance curve). This is the impedance curves
of the UHF dipole shown in Fig. 7.12.
shown in Fig. 7.15. The resistance of this new design has been lowered to well below
20 Ω for the entire UHF band for RFID (860 - 960 MHz).
From Section 6.2, the power loss caused by mismatch can be obtained using:
PlostPavailable
=∣∣∣∣Zant − Z∗
cctZant + Zcct
∣∣∣∣2
(7.1)
where Zant is the input impedance of the antenna and Zcct is the input impedance of
the RFID chip used.
From Section 6.6, at 960 MHz, the input impedance of the C1G1 chip used is ap-
proximately 16 − j140 Ω. To compare with other C1G2 chips, the input impedance
is 10.764− j126.61 Ω for an Alien C1G2 chip, 9.0070− j57.806 Ω for a Texas Instrument
C1G2 chip. The design was carried out to match the C1G1 chip. At this stage the power
loss is computed to be around 55 % using (7.1), contributed by mismatch in both the
real and imaginary parts.
Page 209
7.4 Dual-Frequency Antenna Design
Figure 7.14. A UHF dipole with bent sides. The dipole is bent to conserve space but with slight
reduction in performance.
0.85 0.90 0.95 1.00 1.05 0.00
50.00
100.00
150.00
200.00
Freq [GHz]
Figure 7.15. Impedance of the UHF dipole with bent sides. (Red curve is the resistance curve
and blue curve is the reactance curve, Y-axis in Ω). This is the impedance curves of
the antenna shown in Fig. 7.14.
7.4.5 Compatibility of Dipole in HF
The dipole shown in Fig. 7.14 has a reasonably good match with an RFID UHF chip
within the UHF band (Fig. 7.15). However, the matching network of the UHF dipole
has provided a short circuit path for HF. The initial idea to rectify this problem is to add
a series capacitance which acts like an open circuit at HF and acts like a short circuit
Page 210
Chapter 7 Dual-Frequency RFID Antenna
at UHF. The additional series capacitance takes the form of a gap, with an underlying
track, as shown in Fig. 7.16.
Figure 7.16. A UHF dipole with bent sides and gap. The gap prevents the UHF matching
network from short-circuiting the HF antenna at HF.
To maintain a balanced current flow, it is better to reposition the gap so that the gap
is symmetric along the y-axis. The end result is as shown in Fig. 7.17, (top view in
Fig. 7.18). The patch at the bottom of the gap can be adjusted easily (in both length and
width) for fine-tuning purposes.
Figure 7.17. A UHF dipole with bent sides and repositioned gap. The gap is repositioned to
maintain balanced structure for a balanced current flow.
Page 211
7.4 Dual-Frequency Antenna Design
Figure 7.18. Final version of the UHF dipole, with matching network, bent sides and gap.
The simulation results of the final design of the UHF dipole with matching network
(Fig. 7.18), shown in Fig. 7.19 and Fig. 7.20, show the impedance in the HF band
and UHF band respectively. At HF, the antenna has a relatively large impedance.
Though the real part is significantly smaller than the imaginary part, at 13.56 MHz,
its impedance is 57.4 - 4461.67 Ω. At UHF, its impedance is a bit higher at the 860 MHz
end, but can be adjusted easily, by changing the width of the underpass track below
the gap. As mentioned before, the design is focussed at 960 MHz, which is the highest
end of the RFID operating frequency. At 960 MHz, the simulated impedance is 6.68 +
139.27 Ω.
7.4.6 Merging of the New UHF Dipole with the HF Coil
A dipole which provide a good matching with a UHF RFID chip at UHF is presented in
the previous section. The attention is now to merge the independently designed UHF
dipole and HF coil antenna and to obtain a resonance at HF, while maintaining a good
match at UHF.
There is no available dual-frequency RFID tag chip. It is assumed that a dual-frequency
RFID chip will not provide any input capacitance at HF. It should be pointed out
that this assumption may not be true in practice as most of the available HF chips
Page 212
Chapter 7 Dual-Frequency RFID Antenna
5.00 10.00 15.00 20.00 25.00-12500.00
-10000.00
-7500.00
-5000.00
-2500.00
0.00
2500.00
Freq [MHz]
Figure 7.19. HF Impedance of the final version of the UHF dipole shown in Fig 7.18. (Red
curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).
0.85 0.90 0.95 1.00 1.05 0.00
50.00
100.00
150.00
200.00
250.00
Freq [GHz]
Figure 7.20. UHF Impedance of the final version of the UHF dipole shown in Fig 7.18. (Red
curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).
have an input capacitance of 23.5 pF (Section 6.6). Nonetheless, the (capacitive) input
impedance of an RFID tag chip can be represented by a parallel capacitance and resis-
tor (Section 6.4). The additional of this parallel capacitance can be compensated easily
by reducing the overall capacitance in the HF coil of the dual-frequency antenna.
With the above stated assumption, the dual-frequency antenna itself must have a self-
resonance at HF. The individual HF antenna designed in previous section has a higher
Page 213
7.4 Dual-Frequency Antenna Design
resonant frequency than the target resonant frequency (Fig. 7.4). Following f0 = 12π
√LC
,
where L and C are the inductance and the capacitance of the antenna respectively, ei-
ther L or C or both L and C must be increased to achieve this HF resonance. To increase
L, the number of turns can be increased or the track width can be reduced. To increase
C, the gap between tracks can be reduced.
However, before any fine-tuning was finalised, the UHF dipole must be considered as
well. As mentioned before, the UHF dipole has a gap so that it will look like an open
circuit at HF. This is not exactly true as Fig. 7.19 shows that the impedance of the UHF
dipole at HF is not near infinity.
The UHF dipole will be acting like a capacitor in parallel with the HF loop antenna
(which is inductive overall). However, the inductance and the capacitance are not high
enough to have a resonance point at 13.56 MHz, which is actually a very low frequency
to be achieved with respect to the size of the antenna.
Even with the UHF dipole attached, the four turn HF planar coil antenna is found to
have too high a self-resonance frequency as shown in Fig. 7.21. Hence the first step is to
increase both the inductance, L, and capacitance, C, of the HF antenna. The dimensions
could be modified aggressively to achieve this but there is an easier approach.
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00-20000.00
0.00
20000.00
40000.00
60000.00
Freq [MHz]
Y 1
Figure 7.21. Self resonance frequency of HF planar coil antenna shown in Fig. 7.6. (Red
curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).
Page 214
Chapter 7 Dual-Frequency RFID Antenna
The solution is to have a back-to-back multi-turn loop antenna on the both sides of
the FR4. Not only the length of the coil is doubled (which increases the L value), the
overlapping coil tracks on the both sides of the antenna will provide extra C to the
antenna. Via holes are used to connect the planar tracks on both sides.
Having a back-to-back multi-turn loop antenna also solves the problem in determining
the length of the transmission line, l. Since the back-to-back tracks are acting like a
capacitor, which allows UHF signals to bypass most of the inductive loop, the HF loop
antenna appears to be a short circuit at UHF. Using l equal to a quarter of a wave-
length, the short circuited HF loop antenna is transformed to an open circuit. The
design of the antenna at this stage is shown in Fig. 7.22.
Figure 7.22. Merging of HF multi-turn loop antenna with a UHF dipole. The HF loop antenna
used is the double-sided version of the HF loopantenna shown in Fig. 7.1. The UHF
dipole used is as shown in Fig. 7.18.
It can be seen that the overall dimension in the x-axis is caused by the HF antenna.
Hence the HF antenna is redesigned to reduce the overall dimension. The idea is to
have the number of loops reduced by one. To cater for the reduction in L and C caused
by the reduction of the number of loops, the coil track is made thinner, reduced from
3 mm to 2 mm. The final design is as shown in Fig. 7.23 with all the dimensions.
Fig. 7.28 shows the simulated impedance values of the dual-frequency antenna within
the UHF band. At 960 MHz, the antenna impedance is 24 + 143j Ω, which is very
Page 215
7.4 Dual-Frequency Antenna Design
Via
Via
g
luhf
wuhf
Feed
st
wt_hfwt_uhf
wn
gn
lhf
(a)
whf
lhf
Via
Via
winner
linner
lunder
wt_uhf
(b)
Figure 7.23. Final design for dual-frequency antenna. (a) Front view, lhf = 98; luhf = 43; g =
2; wt hf = 2; wt uhf = 3; st = 6; gn = 3; wn = 16; wuhf = 37. (b) Rear view, linner
= 82, winner = 46; lunder = 38. All units in mm.
near to the intended impedance of 17 + 150j Ω. For HF, since the antenna represents
a parallel resonant circuit, the resistance is theoretically infinite at 13.56 MHz, while
the reactance is 0 Ω. The simulated impedance values of the designed dual-frequency
Page 216
Chapter 7 Dual-Frequency RFID Antenna
4.00 6.00 8.00 10.00 12.00 14.00 16.00-10000.00
0.00
10000.00
20000.00
Freq [MHz]
Figure 7.24. Simulated impedance for the dual-frequency antenna at HF. (Red curve is the
resistance curve and blue curve is the reactance curve, Y-axis in Ω). A resonant point
is observed at approximately 13.56 MHz.
antenna at HF is shown in Fig. 7.24. A resonant point is observed at approximately
13.56 MHz. The matching in UHF impedance and the resonant point at HF shows
that the designed dual-frequency antenna has achieved the design aims shown in Sec-
tion 7.3.
7.5 Antenna Fabrication and Testing
The final design for the dual-frequency antenna (Fig. 7.23) is fabricated on a double-
sided FR4 material (Fig. 7.26) and the actual measurement is taken for both HF and
UHF operation. In the case of HF, the antenna resistance is too high at resonance to
have a proper impedance measurement using a network analyser. The unbalanced
input of the analyser also poses problems of measurement.
To demonstrate the fabricated antenna has a resonant point at HF, a transmission mea-
surement is used, where a wide-band loop antenna is used as the transmitting antenna
and the fabricated antenna is used as the receiving antenna as shown in Fig. 7.25. Ro-
tation of the loop confirms that coupling is via magnetic field. The measurement result
Page 217
7.5 Antenna Fabrication and Testing
is shown in Fig. 7.27. The exact resonance point is at 13.94 MHz, which is very near to
the designed frequency of 13.56 MHZ.
Figure 7.25. A transmission measurement using a wide-band loop antenna. The wide-band
loop antenna is placed near to the designed dual-frequency antenna with the plane of
the loop antenna in parallel with the plane of the dual-frequency antenna. When the
loop antenna is rotated 90 so that the plane of the loop antenna is perpendicular to
the plane of the dual-frequency antenna, no transmission is recorded.
Figure 7.26. Fabricated final version of the designed dual-frequency antenna.
Page 218
Chapter 7 Dual-Frequency RFID Antenna
Figure 7.27. Transmission measurement using network analyser to locate the resonance
point at HF. Best transmission point occurs at 13.9 MHz.
Impedance measurement was carried out from 860 - 960 MHz, and compared with the
simulated results, as shown in Fig. 7.28. It was discovered that the measured resis-
tance is higher than the simulated results, but the measured reactance is lower than
the simulated results. The impedance at 960 MHz is measured to be 56.26 + 134.5 Ω.
This discrepancy is probably due to balanced load measuring problem, where the an-
tenna is a balanced load while the network analyser is unbalanced. This can be solved
by having a RF transformer or by connecting a BALUN to the antenna or employing
other means of suppressing common mode currents. However this will not be dis-
cussed further in this chapter but will be considered in the Chapter 9 of this thesis.
Nonetheless, the measurements have shown that a dual-frequency antenna has been
successfully designed and fabricated, which could be used in an RFID dual-frequency
operation.
So far we have presented a detailed design for a high frequency ratio dual-frequency
antenna. Simulated and measured results are shown to confirm its functionality at
both HF and UHF, as specified in the design aims. This shows that with proper design,
a single feed dual-antenna of very high frequency ratio can be achieved. We haven’t
focussed on the optimisation of the antenna size, which will be presented in the next
two sections.
Page 219
7.6 Miniaturisation of Dual-Frequency RFID Antenna with High Frequency Ratio
0
20
40
60
80
100
120
140
160
180
840 860 880 900 920 940 960 980
Frequency (MHz)
Impedan
ce
(Ohm
)
Simu_Re Simu_Im Meas_Re Mea_Im
Figure 7.28. The simulated and measured results for the frequency response of the fabri-
cated dual-frequency antenna within the RFID UHF band from 860 - 960 MHz.
7.6 Miniaturisation of Dual-Frequency RFID Antenna with
High Frequency Ratio
With a functioning dual-frequency antenna, the attention is now focussed on the de-
sign of a miniaturised dual-frequency Radio Frequency Identification (RFID) antenna,
which can support both the HF (13.56 MHz) and UHF bands (860 - 960 MHz). A
novel dual-frequency RFID antenna with high frequency ratio is presented in Fig. 7.23.
The dimensions of this antenna are 114 mm (l) by 98 mm (w), giving a total area of
11172 (mm)2. It has a resonant point at HF and has an impedance match to an RFID
UHF chip impedance (17-150j Ω at 915 MHz).
In designing the miniaturised version of dual-frequency antenna, some of the charac-
teristics from previous dual-frequency antenna design will be retained:
1. The new design uses a similar idea to obtain a dual-frequency antenna, which is
by designing a HF and a UHF antenna separately. These two antennas are then
merged together with a single feed. The HF antenna used is a HF coil antenna
and has a resonance point at 13.56 MHz. The UHF antenna chosen is a UHF
electric dipole with a matching network so that it has a impedance match with
the chip impedance at the operating frequency.
Page 220
Chapter 7 Dual-Frequency RFID Antenna
2. A transmission line (with length lt in Fig. 7.41(a)) is used to link the HF and the
UHF antenna together with a single feed point, while preventing the HF antenna
from affecting the UHF antenna.
3. It is double sided. The HF coil on the top overlaps with the HF coil on the bottom
to provide enough capacitance in order to obtain a resonant point at HF.
4. A series capacitor is added at the matching network of the UHF antenna (gap pro-
vided by g in Fig. 7.41(a)). This series capacitor will have negligible impedance
at UHF but will have a very high impedance at HF. This is to prevent the UHF
antenna from shorting the HF antenna.
5. The HF antenna will provide a DC path, and this is required in certain rectifier
circuit implementations.
7.7 Improvement on Novel Dual-Frequency Antenna
In order to have a miniaturised version of a dual-frequency RFID tag antenna, some
major changes have been made on top of those retained characteristics mentioned in
Section 7.6. Also, some modifications were carried out to improve the overall perfor-
mance of a dual-frequency RFID tag antenna.
1. Relocation of UHF dipole
This improvement idea is to reposition the UHF dipole inside the HF planar coil
antenna to reduce overall size of the dual-frequency antenna. The model is as
shown in Fig. 7.29. The UHF dipole had to shrink slightly to fit.
The impedance of this antenna is as shown in Fig. 7.30 and Fig. 7.31 for HF
and UHF respectively. The HF frequency is slightly off the target frequency of
13.56 MHz, but this can be fine-tuned reasonably easily. However, the most chal-
lenging is the UHF impedance graph, where the impedances fluctuate as the fre-
quency increases. This means the antenna can only be matched to the chip at a
single frequency point only.
Page 221
7.7 Improvement on Novel Dual-Frequency Antenna
Figure 7.29. A different version of a dual-frequency antenna where the dipole is located in
the inside of the HF coil. This is to reduce the overall area of the dual-frequency
antenna.
4.00 6.00 8.00 10.00 12.00 14.00 16.00-10000.00
0.00
10000.00
20000.00
Freq [MHz]
Figure 7.30. The HF impedance of the dual-frequency antenna in Fig 7.29. (Red curve is the
resistance curve and blue curve is the reactance curve, Y-axis in Ω).
2. Miniaturisation of HF coil
Since in the new design, the overall size of the dual-frequency antenna is deter-
mined by the size of the HF coil, we want a smaller HF coil. The new HF coil
is reduced from 98 mm by 65 mm to 81 mm by 58 mm, the gap is 0.5 mm and
the track width is 1 mm. The width of the transmission line is 6 mm. The rule
Page 222
Chapter 7 Dual-Frequency RFID Antenna
860.00 880.00 900.00 920.00 940.00 960.00-50.00
0.00
50.00
100.00
150.00
200.00
250.00
Freq [MHz]
Figure 7.31. The UHF impedance of the dual-frequency antenna in Fig 7.29. (Red curve is
the resistance curve and blue curve is the reactance curve, Y-axis in Ω).
of thumb in designing an HF coil antenna is to maximise the dimension of the
interior area (winner × linner) to maximise the total inductance (including self in-
ductance and mutual inductance) of the coil. With the reduction of the size of
the HF coil, the problem of lower inductance which results in higher resonance
frequency than the intended frequency will be encountered. This problem can be
solved by careful design and by introducing an additional loop into the HF coil
as shown in Fig. 7.32.
Figure 7.32. A miniaturised HF coil antenna. With reduced track width and spacing.
3. Miniaturisation of UHF dipole
Page 223
7.7 Improvement on Novel Dual-Frequency Antenna
With the size reduction of the HF coil antenna, the UHF dipole antenna can be
reduced as shown in Fig. 7.33. Basically, the size of the UHF antenna together
with its matching network is designed to have a thinner track width, shorter
length along x-axis, but with a longer bent side.
Figure 7.33. A miniaturised UHF dipole. With thinner tracks and longer side bending.
The miniaturised HF coil antenna (Fig. 7.32) and UHF dipole (Fig. 7.33) are then
merged together to obtain a new and miniaturised dual-frequency antenna as
shown in Fig. 7.34.
4. HF not affecting UHF over a broader bandwidth
Initially, it was believed that the HF antenna would behave like a short circuit at
the UHF band as the overlapping HF coil provides a short circuit path for UHF
signal (Refer Section 7.4.3). Hence, it was decided to use a transmission line of
length λ4 to transform this short circuit into an open circuit, so that the HF coil
would not affect the operation of the UHF dipole.
However, it was later discovered that the above statement is only true for a single
frequency point. The HF coil antenna is not exactly a short circuit for the entire
band of UHF as shown in Fig. 7.35(a). If a λ4 transmission line of any characteristic
impedance is simply introduced into the structure, the HF antenna will still affect
the UHF antenna over a large portion of UHF band.
Page 224
Chapter 7 Dual-Frequency RFID Antenna
Figure 7.34. A dual-frequency antenna with miniaturised HF coil and UHF dipole. Formed
using the miniaturised HF coil antenna shown in Fig. 7.32 and the miniaturised UHF
dipole shown in Fig. 7.33
Hence, the idea is to design a transmission line with high characteristic impedance,
Z0. The type of transmission line chosen consists of coplanar strips (CPS). CPS
has a theoretical physical limit on the lowest and highest Z0 of 45 and 280 Ω re-
spectively when εr is 10 [86]. The highest Z0 limit can be increased by reducing
the εr value.
The Z0 of the CPS can be increased by changing wt and st (Refer Fig. 7.41(a)). If
the CPS has a very high Z0, the impedance of the designed antenna can be re-
normalised in Smith Chart (by default normalised by Z0 of 50 Ω) to the left hand
region of the Smith Chart. Then, this impedance can be transformed to the right
hand region using a λ4 long CSP.
In mathematical terms, the reflection coefficient, Γ of an antenna is given by:
Γ =ZL − Z0
ZL + Z0(7.2)
where the characteristic impedance (or normalising impedance), Z0, is purely
resistive. Rearranging (7.2) will give us:
Page 225
7.7 Improvement on Novel Dual-Frequency Antenna
1.0 1.00.0
0
10
20
30
40
50
6070
8090100110
120
130
140
150
160
170
180
-170
-160
-150
-140
-130
-120-110
-100 -90 -80-70
-60
-50
-40
-30
-20
-10
5.002.001.000.500.200.000.000.000.000.000.000.000.000.00
5.00
-5.00
2.00
-2.00
1.00
-1.00
0.50
-0.50
0.20
-0.20
0.0
1
2
(a)
1.0 1.00.0
0
10
20
30
40
50
6070
8090100110
120
130
140
150
160
170
180
-170
-160
-150
-140
-130
-120-110
-100 -90 -80-70
-60
-50
-40
-30
-20
-10
5.002.001.000.500.200.000.000.000.000.000.000.000.000.00
5.00
-5.00
2.00
-2.00
1.00
-1.00
0.50
-0.50
0.20
-0.20
0.0
1
(b)
Figure 7.35. Impedance transformation in Smith Chart where all the traces cover 860 to
960 MHz. (a) Trace 2 is the reflection coefficient of the impedance of the antenna
normalised to 50 Ω while Trace 1 is the reflection coefficient of the impedance of the
antenna re-normalised to 290 Ω. (b) Trace of the reflection coefficient when Trace 1 in
(a) is transformed using a λ4 transmission line. Note that the length of the transmission
line corresponds to λ4 of a particular frequency only (not the entire frequency band).
Hence (b) is not exactly (only similar to) Trace 1 of (a) rotated by 180.
Page 226
Chapter 7 Dual-Frequency RFID Antenna
ZL = Z01 + Γ1 − Γ
(7.3)
And with a new characteristic impedance (of the CPS), the new Γ can be com-
puted using:
Γnew =Z0
1+Γ1−Γ − Znew
0
Z01+Γ1−Γ + Znew
0(7.4)
The effect is shown in Fig. 7.35(a), and with the impedance transformation of theλ4 CPS for the designed frequency (approximately λ
4 for other frequencies within
the band of operation), the HF coil impedance is plotted as Fig. 7.35(b). It can be
seen that the HF coil behaves like a high impedance for most of the UHF band of
interest.
The final design of the transmission line linking the HF and the UHF antenna
together has the following design values (Refer Fig. 7.41(a)): st = 6 mm, wt =
0.8 mm and εr = 4.4, resulting a Znew0 ≈ 290 Ω. The effective dielectric constant
value, εe f f is approximately 2.12.
Hence, the length of the λ4 CSP is:
lt =λ
4 ×√εr
(7.5)
At 920 MHz, λ is approximately 0.326 m. Hence lt is approximately 0.056 m or
56 mm.
5. The coupling effect between UHF and HF antenna
As mentioned before, a UHF electric dipole antenna with matching impedance to
the RFID tag chip was designed and simulated. A transmission line was used to
separate the HF and UHF antennas. In the previous dual-frequency design, the
UHF dipole antenna is located beside the HF coil. Although the overall size is in-
creased, the coupling between the HF and UHF antennas is negligible. In the new
design, it is decided to move the UHF dipole inside the HF coil to reduce over-
all size. However, the coupling between the antennas changes the UHF dipole
Page 227
7.7 Improvement on Novel Dual-Frequency Antenna
impedance quite significantly and the UHF dipole no longer has an impedance
match with the UHF chip at UHF. A fine tuning was carried out by re-adjusting
g, luhf and lunder.
6. Reduction of HF antenna size
To fit both the transmission line and the UHF antenna within the HF coil an-
tenna, the HF coil antenna is rotated 90 degree from the previous design (com-
pare Fig. 7.18 and Fig. 7.36). With the HF coil rotated, the simulation results are
as shown in Fig. 7.37 and Fig. 7.38 for the antenna impedance in HF and UHF
respectively. It can be seen that it is still functioning well as a dual-frequency
antenna.
Figure 7.36. Dual-frequency antenna with rotated HF coil. The rotation of HF coil does not
change much on the characteristics of the dual-frequency antenna.
However, the antenna in Fig. 7.36 does not contribute to the aim of minimising
the size of the antenna. Hence, ultimately, the UHF antenna has to be minimised
and relocated inside the HF coil antenna as discussed before. The impact on the
size limitation of the UHF dipole antenna is significant. The wuhf is shortened
from 95 mm to 44 mm. Although this dipole antenna can be designed to match
in impedance with the RFID chip, its simulated directivity drops from 2.9 to 2.2.
This is a trade-off between size and performance.
Page 228
Chapter 7 Dual-Frequency RFID Antenna
5.00 10.00 15.00 20.00 25.00-20000.00
-10000.00
0.00
10000.00
20000.00
30000.00
Freq [MHz]
1
Figure 7.37. Simulated results at HF for antenna in Fig. 7.36. (Red curve is the resistance
curve and blue curve is the reactance curve, Y-axis in Ω). Point 1 indicates a resonant
point at 12.30 MHz.
860.00 880.00 900.00 920.00 940.00 960.00 0.00
50.00
100.00
150.00
200.00
Freq [MHz]
Figure 7.38. Simulated results at UHF for antenna in Fig. 7.36. (Red curve is the resistance
curve and blue curve is the reactance curve, Y-axis in Ω).
Note that the six improvements presented in this section are not subsequence improve-
ments. For example, the reduction of HF antenna size (improvement 6) does not take
place after all the first five improvements. All these improvements complement each
other and contribute to the final design of the miniaturised dual-frequency RFID tag
antenna. The complete blue print of the final design can be found Fig. 7.41.
Page 229
7.8 Testing on Miniaturised Antenna
A B
Figure 7.39. Figure to illustrate the position of RFID chip attached to the antenna. In
the initial testing both HF and UHF chips are attached to feed point A. In the final
prototype, UHF chip is located at point A while HF chip is located at point B. The
impedance of the transmission line is approximately 290 Ω.
7.8 Testing on Miniaturised Antenna
The prototype of the designed dual-frequency antenna (Fig. 7.41) was tested using the
methods described in Section 7.5, and found to be resonating at 13.56 MHz while hav-
ing reasonable impedance match with a generic UHF RFID tag chip at UHF. Hence the
novel dual-frequency RFID tag antenna shown in Fig. 7.23 has been miniaturised suc-
cessfully while its HF and UHF operational characteristics are preserved in achieving
the design aims presented in Section 7.3.
The next testing procedure is to measure the read range performance of the minia-
turised dual-frequency antenna. The biggest challenge in real life testing for a dual-
frequency RFID tag is that there is no single tag chip available that would work on
both HF and UHF. It is the design aim to have a single feed system. However, for the
purpose of testing, an UHF RFID C1G1 chip and a Tag Talk First HF chip (72128 TTF
chip on smart card module with negligible input capacitance) are used instead.
Referring to Fig. 7.39, both the HF (72128 TTF) and UHF chip (Alien C1G1) are attached
at position A. In theory, the read range of HF operation is limited by the size of the
reader antenna [95]. A quick measurement shows that this dual-frequency RFID tag
offers a maximum possible read range at HF when using a square loop HF reader
antenna with edge-to-edge length of approximately 0.3 m.
Page 230
Chapter 7 Dual-Frequency RFID Antenna
However, the HF chip affects the UHF chip significantly and a poor read range of only
0.3 m is observed at UHF. If the HF chip is moved to the B position in Fig. 7.39, the
HF chip will be transformed to a very high impedance at UHF by the transmission
line. Hence the UHF chip will perform normally and a read range of more than 2 m is
observed. A quick check again shows that this dual-frequency RFID tag still offers a
maximum possible read range at HF.
Although the final prototype dual-frequency RFID tag has two chips and has two feed
points, it is believed that this prototype will work with equivalent efficiency in both
HF and UHF bands when a single dual-frequency chip is used on a single feed.
To compare this miniaturised dual-frequency RFID tag with commercial tags, a com-
mercial HF RFID tag and a commercial UHF RFID tag are placed side by side and
a read range measurement in both the HF and UHF bands is carried out. It is dis-
covered that the read range performances are comparable when the tags are alone or
placed near to each other. The conclusion is that, the prototype dual-frequency an-
tenna antenna is superior to the commercial tags because it can handle operations in
both frequency bands with only one feeding point. More discussions on the dual feed
point dual-frequency antenna can be found in Section 8.5.
7.9 Final Design of Miniaturised Dual-Frequency Antenna
The final design of the miniaturised dual-frequency antenna is as shown in Fig. 7.40.
In short, the new design is to relocate the UHF electric dipole antenna inside the HF
coil antenna to reduce the overall size as shown in Fig. 7.41. The total dimension is
81 mm (lhf) by 58 mm (whf), giving a total area of 4698 (mm)2.
7.10 Conclusion
A novel design for a dual-frequency RFID antenna with high frequency ratio was pre-
sented by the method of merging an HF antenna together with a UHF antenna. This
Page 231
7.10 Conclusion
Figure 7.40. A miniaturised dual-frequency antenna. With HF and UHF RFID chips attached.
Via
Via
Feed A
l
w
l
w
t
t
luhf
wuhf
g
hf
hf
ts
Feed B
(a) Front view
l
innerw
inner
Via
Via
lunder
(b) Back view
Figure 7.41. Final antenna design with design parameters. The final prototype has the following
design values in mm: lhf = 81; whf = 58; linner = 70; winner = 48; luhf = 37; wuhf =
44; g = 3; wt = 0.8; st = 6; lunder = 9.
antenna works well in both the HF and UHF bands for RFID operation with a sin-
gle feed. A functioning dual-frequency RFID tag would prove that having a dual-
frequency RFID system is feasible. Although no dual-frequency RFID chip is available
at the moment, it is hoped that this successful design of a compact dual-frequency
Page 232
Chapter 7 Dual-Frequency RFID Antenna
RFID antenna will catalyse the development of a dual-frequency RFID chip. A minia-
turised version was also presented to demonstrate the potential of this dual-frequency
antenna.
Page 233
Page 234
Chapter 8
Alternative Dual-frequencyAntenna Designs
THISchapter complements the research of a novel dual-frequency
antenna of high frequency ratio presented in previous chapter, by
exploring alternative methods in creating a dual-frequency an-
tenna of similar characteristics.
Page 235
8.1 Introduction
8.1 Introduction
As described in Chapter 7, three methodologies were brainstormed to create a single
feed point dual-frequency RFID antenna which operates in both HF (13.56 MHz) and
UHF (860 - 960 MHz). These three methodologies are:
1. Merging an HF antenna and a UHF antenna.
2. Making an HF antenna that operates at UHF.
3. Making a UHF antenna that operates at HF.
Chapter 7 has presented a fully functional dual-frequency RFID tag antenna with its
miniaturised version, developed based on the first methodology, where individual HF
and UHF antennas are designed separately, and combined in a way where the interfer-
ence between each other are minimised even with only a single feed point.
This chapter will cover the alternative methods of merging an HF antenna and a UHF
antenna, and also the remaining two brainstormed methodologies, which are (1) in
Section 8.3, making an HF antenna that operates at UHF, and (2) in Section 8.4, making
a UHF antenna that operates at HF.
This chapter will also investigate the case where two independent HF and UHF RFID
tag are located near to each other, to create a dual-frequency tag but with two isolated
antenna and two independent chips. This is presented in Section 8.5, followed by in
Section 8.6 the comparisons of all methods presented in this thesis to create a novel
dual-frequency antenna, and in Section 8.7 at the end of this chapter by conclusions.
8.2 Alternate Methods in Merging an HF Antenna and
a UHF Antenna
Two methods different from the approach shown in Chapter 7 are investigated for
merging an HF antenna and an UHF antenna, to create a dual-frequency antenna. The
Page 236
Chapter 8 Alternative Dual-frequency Antenna Designs
first approach is to change the position of the feed point while the second approach
uses different UHF antennas (the UHF antenna used in the design in Chapter 7 is a
UHF dipole antenna) for the merging.
8.2.1 Different Feed Point Location
As discussed in the end of Section 7.4.3 and shown in Fig. 7.9, there are two possible
selections of feed point position, at position A or position B. Also the arguments pre-
sented in that section concluded that a choice of position A may make the fine-tuning
process of a dual-frequency antenna difficult as compared to the choice of placing the
feed point at position B. In this section, the results obtained in the case where the feed
point is repositioned to location A are presented and from the following results it is
concluded that a change in the feed location from location B to location A is not feasi-
ble in creating a dual-frequency antenna.
lA B
Figure 8.1. Feed point repositioning of the dual-frequency antenna. Feed point relocated from
position B to position A.
The dual-frequency antenna of interest is shown in Fig. 7.23. This antenna was chosen
as it has been proven in Chapter 7 as a fully functional dual-frequency antenna at both
HF and UHF bands. In the original form, this antenna has a feed point at position
B. The feed point has been relocated to position A for the analysis presented in this
section.
Simulation was carried out using Ansoft HFSS. The simulation results are as shown in
Fig. 8.2 and Fig. 8.3, showing the impedances of the dual-frequency antenna at HF and
UHF respectively with the new feed point.
Page 237
8.2 Alternate Methods in Merging an HF Antenna and a UHF Antenna
4.00 6.00 8.00 10.00 12.00 14.00 16.00-10000.00
0.00
10000.00
20000.00
Freq [MHz]
Figure 8.2. Impedance of dual-frequency antenna with feed point at position A at HF. (Red
curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω). It can
be seen that a change in feed location does not bring much effect to the impedance of
the antenna at HF.
860.00 880.00 900.00 920.00 940.00 960.000.00
50.00
100.00
150.00
200.00
250.00
300.00
Freq [MHz]
Figure 8.3. Impedance of dual-frequency antenna with feed point at location A at UHF.
(Red curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).
As discussed in Section 7.4.3, when the feed point is chosen at location B, the UHF
dipole will contribute useful additional capacitance to the HF antenna when the merged
antenna is operating at HF. With a change in feed point from position B to A, the
changes in reactance at HF would be negligible. This is proven by Fig. 8.2 where the
resonant point is just below 13.5 MHz, very near to the original 13.56 MHz (Fig. 7.24).
Page 238
Chapter 8 Alternative Dual-frequency Antenna Designs
Note that though there only a negligible change in the resultant reactance, the contri-
bution of the transmission line to the total reactance is in a different form when the
position of feed point was changed from position B to position A. When the feed point
is at position B, the transmission line is contributing parallel capacitance while when
the feed point is at position A, the transmission line is contributing series inductance.
However, the change of feed point brings significant change to the impedance of the
antenna of interest (the dual-frequency antenna in Fig. 7.9 which consists of a UHF
dipole and an HF coil) at UHF. With that change of feed point, the λ4 (UHF) transmis-
sion line is not functioning as intended, where it is supposed to transform the HF coil
impedance into a relatively high impedance at UHF. Now the impedance at UHF of the
UHF dipole is transformed to an undesirable value while the HF coil may be acting as
the main radiating body at UHF.
A quick fix might be thought to be is to remove the λ4 (UHF) transmission line com-
pletely so that the impedance of the UHF antenna is not transformed. However, there
is then no way to avoid the impedance of the HF antenna from disturbing the input
impedance at the feed point.
8.2.2 Different UHF Antenna Types
Section 7.4.1 presented the individually designed HF antenna and UHF antenna, before
they are merged together using a transmission line to have a single feed point. The
chosen HF antenna is a multi-turn planar coil antenna, while the chosen UHF antenna
is an electric dipole. This section will explore the alternative antennas to be used as the
UHF antenna.
1. UHF Patch Antenna
A simple UHF patch antenna is as shown in Fig. 8.4. The patch has a width of
99 mm and length 77 mm with the ground plane width and length of 120 mm and
100 mm respectively. The dielectric is a 1.6 mm FR4 with dielectric permittivity
of 4.4.
Page 239
8.2 Alternate Methods in Merging an HF Antenna and a UHF Antenna
Figure 8.4. A UHF patch antenna.
From simulation, the best match point occurs at approximately 880 MHz with
an input impedance of 15 + j70 Ω. It is not a perfect match with any of the UHF
chips shown in Section 6.6. However, from experience, with minimal fine-tuning,
this antenna is definitely an acceptable candidate to be used as a UHF RFID tag
antenna.
Also from simulation, this UHF patch antenna is capacitive at HF. This is a de-
sirable result as discussed in Section 7.4.6, where the UHF antenna is responsible
to provide extra capacitance to the HF coil so a resonance at 13.56 MHz can be
achieved.
Hence, an alternative way to create a dual-frequency antenna is to swap the UHF
electric dipole with a UHF patch antenna. The short-coming of this method is
that the UHF patch antenna cannot be positioned within a HF coil antenna as
compared to the case where an electric dipole can be relocated within a HF coil
antenna (Fig. 7.40). This will unavoidably increase the overall size of the dual-
frequency antenna as the total area is mostly contributed by the area occupied by
both the UHF patch antenna and the HF coil antenna.
2. UHF Loop Antenna
Page 240
Chapter 8 Alternative Dual-frequency Antenna Designs
A UHF RFID tag antenna in the form of loop antenna has been designed and
presented in Section 6.11.3. This loop antenna has a read range of under 1.5 m.
With the inclusion of matching network, this UHF loop antenna is capacitive at
HF. Hence, it can be used to be merged with an HF coil antenna to create a dual-
frequency antenna.
In short, any UHF antenna with matching impedance at UHF with respect to a UHF
RFID chip, and is capacitive at HF, can be used as the UHF antenna to be merged
with an HF coil antenna to form a dual-frequency antenna, using the merging tech-
nique where a HF antenna and a UHF antenna are joined by a λ4 transmission line
(Section 7.4.6). If a UHF antenna is not capacitive at HF, a series capacitor can be added
to achieve this.
8.3 HF Antenna Acting as UHF Antenna at UHF
This section investigates the performance of an HF antenna at UHF. The aim is to de-
sign an HF antenna which operates at HF at the beginning, and with proper tuning
later, it is hoped that this same HF antenna will also have a good match with a UHF
chip. If this is the case, this HF antenna will be acting as a UHF antenna at UHF, and
this HF antenna is in fact a fully functional dual-frequency antenna.
The antenna of interest is an un-tuned HF antenna (Fig. 8.5), and this antenna is de-
signed to resonate in the HF range between 20 MHz to 30 MHz.
The simulated impedance of this un-tuned HF antenna is as shown in Fig. 8.6. From the
figure, it can be seen that the resonance frequency is approximately 25.4 MHz, which is
higher than the desired frequency. However, the focus here is not to make this simple
HF antenna to resonate at 13.56 MHz, though, it can be done by adjusting the parallel
capacitance and choosing the right HF chip as shown in Section 6.10.4.
As mentioned before, the aim in this section is to examine whether it is feasible to
have an HF antenna operating at UHF. The simulation results for the same HF antenna
(Fig 8.5) at UHF are as shown in Fig. 8.7.
Page 241
8.3 HF Antenna Acting as UHF Antenna at UHF
Figure 8.5. The HF antenna to be fine-tuned to operate at UHF. The feed point of the antenna
is represented by a small rectangular box. The red colour track is the top coil while
the gray colour track is the overlapping coil on the other side of the dielectric material
(FR4). This HF coil antenna is the early version of HF antenna designed for pig tagging
purposes (Section 6.10.4).
15.00 17.50 20.00 22.50 25.00 27.50 30.00-10000.00
0.00
10000.00
20000.00
Freq [MHz]
Figure 8.6. Simulated impedance of the HF antenna shown in Fig. 8.5 at HF. (Red curve
is the resistance curve and blue curve is the reactance curve, Y-axis in Ω). Resonant
point at 25.4 MHz.
The impedance around 800 MHz to 850 MHz is almost the intended matching impedance
with respect to a RFID tag chip (Best match point at approximately 828 MHz). With
Page 242
Chapter 8 Alternative Dual-frequency Antenna Designs
750.00 770.00 790.00 810.00 830.00 850.00 0.00
50.00
100.00
150.00
200.00
Freq [MHz]
1
2
3
4
Figure 8.7. Simulated impedance of the HF antenna shown in Fig. 8.5 at UHF. (Red curve is
the resistance curve and blue curve is the reactance curve, Y-axis in Ω). Points 1 and
2 indicates the best match point at approximately 828 MHz.
proper tuning, it is confirmed through simulation that this HF antenna can indeed
match in impedance to a UHF chip.
Despite the fact that an HF antenna can be matched to a UHF RFID chip, there remain
a few problems that need to be looked into. Firstly, this antenna is difficult to be fine-
tuned to operate at both HF and UHF. For example, from experience, if the antenna is
tuned to resonate at HF, it will become unmatched in UHF, and vice versa.
Secondly, a normal HF coil has a multi-turn structure to provide sufficient inductances
to obtain resonance at 13.56 MHz. The multi-turn structure is relatively long in terms
of UHF wavelength, and hence even with the impedance matched, it is suspected that
it will be too lossy to be used as a UHF radiator. Evidence on this matter will be
presented in the next section.
Page 243
8.3 HF Antenna Acting as UHF Antenna at UHF
8.3.1 Loss in Antenna
The loss of an antenna can be estimated using the simulated directivity and gain. In
simple terms, gain is the directivity of an antenna, taking into consideration the dissi-
pative loss of an antenna (a thorough discussion of directivity and gain can be found
in Section 6.3).
For the antenna shown in Fig. 8.5 with the feed point indicated by a rectangle in the
same figure, the peak directivity obtained through simulation is 3.5476 but its peak
gain is 0.00046716. This proves that this antenna is a very poor radiator. The following
discussion explains why.
The plot of the current density at frequency 910 MHz is included here as Fig. 8.8. As in
Fig. 8.5, the feed point is represented by a small rectangle in the figure. This antenna
(from Fig. 8.5) has two sides of multi-turn loops. The top side loops are not obscured
by the bottom side loops, however, the bottom side loops (which are gray) are hardly
seen in Fig. 8.8 as they overlap with the top side loops. For the top side loops only, the
current density is shown at a particular point in time using the colours indicated in the
legend.
Figure 8.8. The structure of the HF antenna shown in Fig. 8.5 with illustration of surface
current density on the top side. The feed point is represented by a small rectangular
box.
Page 244
Chapter 8 Alternative Dual-frequency Antenna Designs
Upon a careful inspection, it can be seen that there is a standing wave pattern in the
plot of current density. We conclude that this antenna has too long a track for operation
at 910 MHz. Current changes phase along the multi-turn coil, with approximately a
phase change every turn. Radiation from currents flowing in adjacent turns will then
tend to cancel.
Though the current density shown in Fig. 8.8 is a just snapshot of the current density,
this figure shows that the radiation properties of this antenna are hard to control at the
frequency of interest.
8.3.2 Antenna Measurement Results
As discussed in previous section, it is highly suspected that insufficient power will
reach the feed point of the antenna when the antenna is receiving.
To confirm the claim that the HF antenna is too lossy to be used as a UHF antenna,
the HF antenna shown in Fig. 8.5 is fabricated with a UHF chip is attached to it. A
quick measurement test was devised. In this test, a small wide-band loop antenna
(with diameter of approximately 20 mm), was connected to a network analyser. With
the network analyser measuring the reflection at the port connected with that loop
antenna, the HF antenna with a UHF chip was positioned near that loop antenna. The
reflection measurement provided a good estimation on how efficiently the HF antenna
was acting as a UHF absorber. The results is shown in Fig. 8.9. A measurement of
< -3 dB reflection (with output power of network analyser set to -20 dBm) shows that
sufficient energy has been absorbed by the antenna for a normal UHF tag operation.
However, when the tag was positioned within the interrogation zone of a UHF reader,
the tag cannot be detected. The distance between the tag and the reader antenna was
reduced slowly towards zero, but no tag detection was recorded by the reader. As
mentioned before, this is most likely caused by the lossy nature of the HF antenna at
UHF and all the energy absorbed by the antenna has been dissipated in the coil, rather
than transferred into the UHF chip.
Page 245
8.4 UHF Antenna Acting as an HF Antenna at HF
Figure 8.9. Measurement result of the HF antenna shown in Fig. 8.5 at UHF. The measure-
ment result is identical to the frequency response of a functioning UHF tag.
This concludes the experiment to make use of an HF antenna as a UHF antenna at UHF.
The truth seems to be an HF antenna, which has a size feasible for RFID application
and to resonate at 13.56 MHz, will be lossy or have an unsuitable current distribution
(not reaching the feed point) at UHF. Even though it can be fine-tuned to match in
impedance to a UHF chip, the losses in the coil or the unsuitable current distribution
prevent sufficient power from reaching the UHF chip.
8.4 UHF Antenna Acting as an HF Antenna at HF
This is the least explored methodology. The reason is that under normal operating
conditions a UHF antenna would not have high net inductance value at HF. Two UHF
antennas are considered, namely the UHF dipole and the UHF patch antenna. From
previous discussions, it has already been shown that all these two UHF antennas are
capacitive at HF. Hence, it is concluded that without an additional matching circuit,
a common UHF RFID tag antenna would not be suitable to be used as an RFID HF
tag antenna. The inclusion of matching network is not a solution as it will de-tune the
UHF antenna and the matching between the UHF antenna and the UHF chip will be
Page 246
Chapter 8 Alternative Dual-frequency Antenna Designs
affected. Also, if a coil UHF antenna (with few turns) is used, a matching over too large
an impedance ratio is required, if this coil antenna is to be tuned to operate at HF.
8.5 Multi-Feed Point Antenna
Although it is the aim to have a single feed dual-frequency antenna, it is of great in-
terest to examine what will happen if one HF and one UHF antenna are placed near to
each other, with two separate chips.
The advantages of having two separate antennas are:
1. HF and UHF tags are readily available. No research or development is required.
2. Modification of the chips, which is necessary for a single feed dual-frequency
antenna, is avoided.
3. Even if redesign of HF and UHF antennas is involved, it is not as complex as the
design of a single dual-frequency antenna.
On the other hand, the disadvantages of this configuration are:
1. The need to program the HF tag and UHF tag to have a same EPC, or a database
is required to link the HF and UHF EPC together.
2. Two chips are required as compared to one needed in a single feed dual-frequency
antenna.
3. The configuration may attract a higher price, as it is basically two tags as com-
pared to one tag based on one single feed dual-frequency antenna.
8.5.1 Stacking UHF and HF Antennas
The idea is to stack a UHF antenna on top of a HF antenna, reducing the horizontal area
needed, but increasing the vertical volume at the same time. This design is based on the
HF pig tag presented in Section 6.10.4 and the UHF pig tag discussed in Section 6.11.4.
The reasons these two antennas are chosen are:
Page 247
8.5 Multi-Feed Point Antenna
• The sizes are almost the same.
Since both the HF and UHF RFID tags designed for pig tagging use the same
encapsulation casing (Fig. 6.29), they are of similar size.
• There is an opening in the middle of the chosen UHF antenna.
An HF antenna will only work when there exists a magnetic field which couples
through the HF coil. To make this possible, the chosen UHF antenna must have
an opening in the middle.
• These antennas can be readily encapsulated using the casing shown in Fig. 6.29.
It is measured that the maximum spacing between the two antennas is 2 mm so
that they can be fitted into the casing.
Fig. 8.10 shows the suggested configuration. The separation distance between the HF
and UHF RFID tags, t, is controlled using small pieces of thin paper.
t
UHF
HF
Figure 8.10. Stacking of HF and UHF RFID tags. The HF and UHF RFID tags are separated
a distance, t, away. Pieces of thin paper are inserted or removed between the two
antennas to control the spacing.
For the HF operation, as long as the UHF antenna is not touching the HF antenna (no
electric conduction), the HF tag can be detected by a HF reader. However, this is not
found to be the case for the UHF tag where the HF tag has a great impact on the read
range.
Read range measurements were carried out while the spacing between these two an-
tenna was varied and the results are recorded in Table 8.1.
When the two tags are too close to each other, the UHF tag was de-tuned and could not
be detected. To re-tune the tag using a network analyser is unexpectedly complex as
Page 248
Chapter 8 Alternative Dual-frequency Antenna Designs
Table 8.1. Measurement results for the UHF read range of the dual-frequency antenna using
the stacking method.
Separation (mm) Read Range (m)
∞ 0.75
10 0.38
9 0.35
5 0.30
the HF tag has some resonant points in the UHF band of interest, as shown in Fig. 8.11.
When the HF tag and UHF tag are placed near to each other, the response looks like
Fig. 8.12, where it is impossible to distinguish which is the UHF response. Fig. 8.13
presents the frequency response when the UHF tag is 5 mm away from HF tag, and
can be interrogated by a reader.
Figure 8.11. Frequency response of the HF tag at UHF without the presence of any UHF
tag. It can be seen that there are several dips in the UHF band caused by an HF tag.
Page 249
8.5 Multi-Feed Point Antenna
Figure 8.12. Frequency response of the UHF tag at UHF when placed near to HF tag. The
distance between the HF and UHF tags is approximately 2 mm. As it can be seen on
the above measurement results, it is very difficult to distinguish the frequency response
of the UHF tag from the HF tag for fine-tuning purposes.
Figure 8.13. Frequency response of the UHF tag at UHF when the tag is able to be read
by an interrogator. The separation between the HF and UHF tag is 5 mm.
Page 250
Chapter 8 Alternative Dual-frequency Antenna Designs
The results showed that a separation between the UHF and the HF antenna is required
for the UHF tag to operate. Nonetheless, the method of stacking can be used to create
a dual feed point dual-frequency RFID tag.
8.5.2 Side-by-side HF and UHF Antenna
In this section, two RFID tags (one HF and one UHF) are placed side-by-side. A com-
mercial UHF tag is often an electric dipole with a matching circuit and is long and thin.
A commercial HF tag on the other hand is always a multi-turn coil antenna. The area in
the middle of the multi-turn coils is a major factor in controlling the total inductance.
A quick test proved that above the HF and UHF tags can be read by HF and UHF RFID
readers as long as the two tags do not overlap each other.
8.6 Comparison Between Different Methodologies
This section compares all the methods presented in this thesis which were used to
develop a dual-frequency RFID tag antenna. Some methods yield better results than
the others.
Table 8.2 summarises the observations on simulation and measurement results. The
best results are from the merging of a UHF antenna with an HF antenna. The dual-
frequency antenna produced using this method, with the UHF antenna being a UHF
dipole and the HF antenna being an HF coil, was presented in Chapter 7.
The choice of the antenna types in using the merging method in producing a dual-
frequency antenna is important. for example, merging an HF coil antenna with a UHF
patch antenna yields the same satisfactory performance, but requires a bigger area
(Section. 8.2.2) while merging an HF coil antenna with a UHF coil antenna produces a
satisfactory dual-frequency antenna (Section 8.2.2).
Page 251
8.6 Comparison Between Different MethodologiesTable
8.2
.Com
par
ison
bet
wee
ndiff
eren
tm
ethodolo
gie
sin
crea
ting
adual-freq
uen
cyRFID
tag
ante
nna.
The
Ref
eren
ceco
lum
nsh
ows
the
loca
tion
inth
isth
esis
wher
eth
ediscu
ssio
nof
diff
eren
tm
ethodol
ogie
sca
nbe
found.
The
Fee
dPoi
nt
Tot
alco
lum
nsh
ows
the
num
ber
of
feed
poi
nt
required
while
the
Fee
dPoi
nt
Loca
tion
isth
epos
itio
nof
feed
poi
nt
with
resp
ect
toth
eco
nfigu
ration
show
nin
Fig
.7.
9.T
he
HF
and
UH
Fco
lum
ns
indic
ate
whet
her
ornot
the
des
igned
dual
-fre
quen
cyan
tenna
isab
leto
oper
ate
inH
Fan
dU
HF
ban
ds
resp
ective
ly.
The
aste
risk
sign
,∗ ,
indic
ates
ades
igned
dual
-fre
quen
cyan
tenna
isei
ther
wor
king
inth
eH
For
the
UH
Fban
d,but
not
inbot
hban
ds
as
inte
nded
.
Met
hodo
logi
esR
efer
ence
Feed
Poin
tH
FU
HF
Tota
lLo
cati
on
Mer
ging
(UH
Fdi
pole
and
HF
coil)
Cha
pter
71
B
Mer
ging
(UH
Fpa
tch
and
HF
coil)
Sect
ion.
8.2.
21
B
Mer
ging
(UH
Fco
ilan
dH
Fco
il)Se
ctio
n8.
2.2
1B
Mer
ging
(UH
Fco
ilan
dH
Fco
il)Se
ctio
n8.
2.1
1A
∗
∗
HF
asU
HF
ante
nna
atU
HF
Sect
ion
8.3
1N
/A
×
UH
Fas
HF
ante
nna
atH
FSe
ctio
n8.
41
N/A
×
Stac
king
ofH
Fan
dU
HF
Sect
ion
8.5.
12
N/A
Side
-by-
side
HF
and
UH
FSe
ctio
n8.
5.2
2N
/A
Page 252
Chapter 8 Alternative Dual-frequency Antenna Designs
Changing the feed point location from position B to position A (Fig. 8.1) is not an option
as the HF antenna and the UHF antenna will interfere each other (Section 8.2.1). The
end result is either the dual-frequency antenna works at the HF band only or at the
UHF band only.
It is also proven that an HF antenna will not be an efficient radiator at UHF (Sec-
tion 8.3), and a UHF antenna will not be able to provide sufficient inductance for HF
chip at HF (Section 8.4).
Using a multi-feed structure is a viable option (Section 8.5.1 and Section 8.5.2). The
only short-coming is that two chips are needed instead of one. However, it is the design
aim to have just a single feed dual-frequency antenna to reduce the cost. Hence, the
best design as recommended by this thesis is the merging of UHF dipole and HF coil
antenna to form a single feed high frequency ratio dual-frequency RFID tag antenna.
8.7 Conclusion
This chapter has offered detailed discussion on all the methods brainstormed in cre-
ating a dual-frequency antenna apart from the most successful method which has al-
ready been presented in a previous chapter. A comprehensive comparison between the
strengths and weaknesses of each method is also included in this chapter. This shows
that there are some flexibilities in designing dual-frequency antennas depending on
the application. This is important as a wide deployment may require various forms of
dual-frequency antenna.
Page 253
Page 254
Chapter 9
Measurement Techniques
THE performance of a passive Radio Frequency Identification
(RFID) tag depends on the antenna and the chip efficiency of the
tag, and also depends on the matching between the antenna and
the chip. The design of the tag antenna often involves the measurement
of its impedance. This chapter investigates the RF cable effect on the mea-
surement of RFID tag antenna impedance. Measurement results obtained
through several common methods are compared with the simulated results.
Ways of minimising measurement error are discussed and an analysis of the
effect of mismatch between the tag antenna and the chip is presented.
Page 255
9.1 Introduction
9.1 Introduction
Various RFID tag antenna designs for HF and UHF RFID tags have been presented
from Chapter 6 to Chapter 8, with simulation and measurement results presented. The
simulation of an RFID tag antenna was carried out using Ansoft HFSS as presented in
Section 6.9.
During simulation, fine-tuning can be carried out to adjust antenna performance pa-
rameters, such as the operating frequency, bandwidth and impedance, to the desired
values. Most of the time, with the restriction of antenna size and cost, the design-
ing stage often involves trade-offs between some of the antenna performance param-
eters. The designed antenna is then fabricated. The biggest challenge is to discover
how a prototype tag antenna can be tested to confirm that it functions as designed,
before large scale production is carried out. Hence, this chapter aims to investigate
the challenge of measuring a small RFID tag antenna using several different methods.
Suggestions are presented to improve the measurement results to reflect the actual
characteristic of an RFID tag antenna.
Though this chapter discusses the measurement techniques devised to obtain the in-
put impedance of the prototype RFID tag antennas, the focus of this chapter is on the
measurement techniques for UHF RFID tag antennas. The measurement of HF RFID
tag antenna is relatively simple for the following reasons:
• HF (13.56 MHz) is relatively low in frequency as compared to UHF (860 - 960 MHz).
It is a known fact that high frequency measurements are more challenging than
low frequency measurements.
• An HF tag antenna is normally low in resistance (at negligible level) and is induc-
tive (a multi-turn coil antenna). A direct measurement using a network analyser
will give a good approximation of the resultant reactance value at 13.56 MHz.
Alternatively, a known value capacitor can be attached across the feed points of
the HF antenna, and by exciting the HF antenna while monitoring the network
analyser, the resonant frequency and the resultant inductance value of the HF
antenna can be obtained. Though, the stated methods are not ideal, as there are
Page 256
Chapter 9 Measurement Techniques
several other problems such as the problem of a balanced load connected to un-
balanced measuring device, experience shows that the impact is not severe, and
the inaccuracy caused by those problems can be simply rectified by fine-tuning.
• Both the resistance and reactance of a UHF RFID antenna are essential in deter-
mining the power transferred from the antenna into the attached chip. Hence a
good measurement of both resistance and reactance is required.
As discussed in Section 6.2, to enable maximum power transfer from a UHF RFID tag
antenna to a UHF RFID tag chip, the tag antenna should have an impedance conjugate
of the input impedance of the tag chip. Also, a tag antenna should have at least unity
gain with a wide antenna gain pattern, to enable easy absorbtion of RF energy from
any RFID reader, and transmission of that energy to the feed point.
There is no literature which explicitly discusses the measurement methods used to
determine the performance of an RFID tag antenna. The closest literature is on the
measurement of small antennas used by a mobile handset, which literature will be dis-
cussed in detail in the next section. Section 9.3 shows the experimental settings while
Section 9.4 presents the experimental results, together with some interpretation of the
results. Further modifications and improvements on measurement were carried out
and discussed in Section 9.5 while the effect of inaccuracy in measurement is presented
in Section 9.6. A conclusion is presented in Section 9.7.
This chapter is the extended version of a published paper [128].
9.2 Background and Literature Review
For a direct feed measurement using a network analyser, an RF cable is required to link
the antenna of interest to the network analyser. There is no doubt that an RF feed cable
will affect the measurement of input impedance, current distribution, and radiation
characteristic of the antenna.
Page 257
9.2 Background and Literature Review
The source of the error is the appearance of common mode current on the outside sur-
face of the outer conductor of a coaxial cable (Fig. 9.1(a)). In the case of antenna mea-
surement, when the antenna is a balanced load, this common mode current is caused
by the balanced to unbalanced connection between an antenna under test (AUT) and
the coaxial cable as shown in Fig. 9.1(b), which is the case in most of the measurements
carried out on the antenna prototypes presented in this thesis where the AUTs are of
balanced structure.
S11 will be the only antenna parameters to be measured and compared with the simu-
lation results in this chapter. Another useful antenna parameter for which comparison
between simulation and experiment may be of interest is the radiation pattern as dis-
cussed in [129]. However, we believe that S11 is more important than radiation pattern
in the case of an RFID tag antenna as the value of S11 allows the evaluation of the
power transfer from the antenna to the tag chip, and power transfer from the antenna
to the tag chip is directly proportional to the read range of an RFID tag. Although it
is good to have the verification of simulated radiation pattern, a tag antenna gain is
averagely unity in all direction and the orientation of a tag antenna with respect to an
RFID reader antenna is random in real life. This will reduce the importance of radiation
pattern of an RFID tag antenna in predicting its read range.
I
I
I
I
I’
I’
(a)
V
(b)
Figure 9.1. Explanation of common mode current. (a) Common mode current flows on the
outside surface of the outer conductor of a coaxial cable (b) Common mode current
caused by the direct connection of a balanced load to an unbalanced connecting cable.
DeMarinis [130] proposed the use of a ferrite choke as an effective solution while Icheln
[131, 132] proposed the use of Balun to transfer the unbalanced nature of a coaxial
Page 258
Chapter 9 Measurement Techniques
cable to a balanced antenna under test (AUT) to minimise the effect of the RF cable.
However, Icheln stressed that the use of Balun outperformed the use of ferrite choke
despite the fact that Balun is often band limited. The reason given was that ferrite
chokes are often lossy and hence affect the measurement of the radiation pattern, which
a properly designed Balun does not.
Alternative methods include a measurement scheme using a microstrip transmission-
line to connect a coaxial cable to an AUT as suggested by Chen [133], and also, the
use of a stack of metal rings to suppress the appearance of common mode current as
presented by Kahng [134].
9.3 Settings and Connection
An RFID tag antenna is normally made of thin copper tracks on an adhesive sheet.
This is to make an RFID tag easily and readily attached to an object. However, for
convenience, the prototype tag antennas under test are fabricated on a solid dielectric
material to have a solid foundation. The material chosen is FR4. An SMA connector
(sometimes with reduced edges for the reason explained later in this section) is then
used as the feed point of the AUT in order to connect the AUT to a network analyser
for measurement. Measurement results are then compared with the simulated results
obtained through Ansoft HFSS simulation software. In all cases, the simulation results
are used as the standard results.
Three types of coaxial cable configuration as shown in Fig. 9.2 were used for testing
purposes:
1. RG223 coaxial cable.
2. RG223 coaxial cable with ferrite sleeve.
3. Gore GMCA 190-1 (now known as G2) coaxial cable.
Two types of antenna were tested using some or all of the coaxial configurations stated
above:
Page 259
9.3 Settings and Connection
1
2
3
Figure 9.2. Cables used for measuring experiment. (1) RG223 coaxial cable. (2) RG223 coaxial
cable with ferrite sleeve. (3) Gore GMCA 190-1 coaxial cable.
1. A balanced bow tie antenna.
A balanced bow tie antenna is fabricated and is shown in Fig. 9.3. The feed point
of the antenna (with an SMA connector) is from the back and through the FR4
material as shown in Fig 9.3 (a).
(a) Picture of the bow tie
antenna
Side View(i) (ii)
Front View
Copper
FR4
SMA
ReducedEdgesVia
Holes
(b) Details and side view of the bow tie antenna
Figure 9.3. A bow tie antenna on FR4. (a) A picture showing the bow tie antenna used in
measurement experiments. (b) Front view and side view of the bow tie antenna with
(i) SMA connector (ii) modified SMA connector to enhance measurement results.
Page 260
Chapter 9 Measurement Techniques
2. Half of the balanced bow tie antenna on ground plane.
The dimensions of the balanced bow tie antenna (Fig. 9.3) are made used of in
fabricating this special antenna, which is half of the balanced bow tie antenna
on a ground plane (Fig. 9.4). This unbalanced version of the bow tie antenna is
soldered on an SMA connector, which is mounted on a ground plane. The idea is
to combine the unbalanced bow tie antenna with its image created by the ground
plane, to simulate a complete balanced bow tie antenna. It has been proven in
theory that a half unbalanced bow tie antenna on a ground plane will have half
of the input impedance of a complete balanced bow tie antenna.
(a) Picture of the half bow tie
antenna
(b) Details and side view of the half bow
tie antenna
Figure 9.4. Half bow tie antenna on ground plane. (a) A picture showing the half bow tie
antenna on ground plane used in measurement experiments. (b) Side view of the half
bow tie antenna on ground plane. The SMA connector does not need to modified
(edged) like the one shown in Fig. 9.3(b) as it is completely shielded by the metallic
ground plane.
The initial measurement results to illustrate the effect of the SMA connector of the
balanced bow tie antenna are shown in Fig. 9.5, where a comparison between the sim-
ulated and measured reactance of a simple bow tie antenna is made.
These initial measurement results show that the SMA feed arrangement of the balanced
bow tie antenna has a significant impact on the results. The size of the SMA connector
is not negligible in relation to the size of the AUT. The edges of this SMA connector
provide an alternate path for displacement current. The solution is to minimise the
edge of the SMA connector as shown in Fig 9.3 (b).
Page 261
9.4 Measurement Results
Also, from Fig. 9.5, it is shown that at a higher frequency, the accuracy of the mea-
surement of reactance drops. This is consistent with the original expectation, where an
accurate measurement is harder to obtain with the increment of frequency.
To avoid the influence of the SMA connector on the measuring results (as the main aim
is to investigate the effect of connecting cables), in this chapter, all the measurements
were carried out with a modified SMA connector, unless the AUT is the half bow tie
antenna on ground plane or otherwise stated.
Comparison of Simulated and Measured Reactance Value
-40
-20
0
20
40
60
80
100
1.1 1.25 1.4
Frequency (GHz)
Re
ac
tan
ce
(O
hm
)
Simulated
Modified SMA
SMA
Figure 9.5. The effect of SMA connector on measurement results. The edges of the SMA
connector increase the capacitance value and hence lower the overall reactance values.
The modified (edged) SMA connector improves the measurement results significantly.
Cable (1) in Fig. 9.2 was used for all these cases.
9.4 Measurement Results
9.4.1 A Balanced Bow Tie Antenna
As mentioned before, simulation results are used as the standard results. The inaccu-
racy of measurements using Cable (1), Cable (2) and Cable (3) are on average similar
as shown in Fig. 9.6 and Fig. 9.7. However, just by touching the cables with a person’s
Page 262
Chapter 9 Measurement Techniques
hand and monitoring the impedance of the AUT, it is confirmed that Cable (1) and Ca-
ble (3), without the addition of ferrite sleeves, are more susceptible to the environment.
Resistance Comparison
14
16
18
20
22
24
26
28
30
860 870 880 890 900 910 920 930 940 950 960
Frequency (MHz)
Resis
tan
ce (
Oh
m)
Simulated
Cable (1)
Cable (2)
Cable (3)
With Ground Plane
With Balun
Figure 9.6. Simulated and measured resistance of AUT. Simulation results are used as the
standard results.
Reactance Comparison
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
860 870 880 890 900 910 920 930 940 950 960
Frequency (MHz)
Reacta
nce (
Oh
m) Simulated
Cable (1)
Cable (2)
Cable (3)
With Ground Plane
With Balun
Figure 9.7. Simulated and measured reactance of AUT. Simulation results are used as the
standard results.
Page 263
9.4 Measurement Results
9.4.2 Half Bow Tie Antenna on Ground Plane
It is observed that the ground plane which the half bow tie antenna is mounted on
provided sufficient shielding between the cable and the AUT. The results obtained
using any of the cable shows negligible difference.
In all comparisons and graphs presented in this section, measured values for a half bal-
anced bow tie antenna on ground plane have been doubled for the reason explained
in the previous section (Section 9.3) to enable easy comparison. In comparison of the
resistance values in Fig. 9.6, the measured results are constantly higher than the sim-
ulated results. This is most probably due to the loading effect of the ground plane.
Nonetheless, in comparison of the reactance values in Fig. 9.7, the measured results
are slightly better than all the results obtained on a balanced bow tie antenna, with
the exception of Cable(3), which is a high quality shielded coaxial cable. Hence, the
ground plane has effectively shielded the cable and the measuring instrument from
any unwanted coupling.
The challenge in using this method is that the AUT must be symmetrical in nature,
which is not always the case for an RFID tag antenna. Also, the distance of the half
bow tie antenna from the ground plane must be half of the distance between the two
feed points of a balanced bow tie antenna.
9.4.3 Results Interpretation
In the real part (resistance) comparison in Fig. 9.6, the half bow tie antenna on ground
plane configuration is the only measurement that gives a higher resistance as compared
to other measurement configurations when compared with the simulation results. This
is probably caused by the loading effect of the finite metal ground plane.
For the imaginary part (reactance), All measurement configurations matched well in
the region 900 to 920 MHz, where the reactance values fall within a 5 Ω range.
The position of ferrite beads affects the measurement results according to [135]. A
small test showed that ferrite beads should not be located too near to the feed point of
Page 264
Chapter 9 Measurement Techniques
the antenna. It will increase the measured resistance, as the ferrite beads will act as an
additional load.
Proper usage of a ferrite sleeve cable (Cable (2)), improves the measurement of resis-
tance quite significantly. However, there is not much improvement in the measurement
of reactance using this method.
9.5 Discussion and Improvement
Using coaxial cable with high shielding can improve the measurement results by re-
ducing the coupling between the cable and the AUT. However, not all errors can
be eliminated as inaccuracy may be introduced by the balanced to unbalanced is-
sue. Hence a Balun was designed by using coplanar strips with an RF transformer
(Fig. 9.8(a)). The model of the chosen commercial RF transformer is TC1-1-13M with
with characteristic impedance of 50 Ω and with operating frequency range of 4.5 to
3000 MHz.
The biggest challenge in using an RF transformer is to have 50 Ω coplanar strips. Ac-
cording to [86], the impedance of coplanar strips is physically limited to above 45 Ω.
Even for 50 Ω coplanar strips, the fabrication process is extremely difficult. Since the
AUT is small, the track size W1 and W2 cannot be too wide (W1 and W2 < 5 mm). To
obtain 50 Ω coplanar strips with W1 and W2 less than 5 mm, the width of separation,
S, has to be around 0.05 mm for FR4 material with εr equals 4.4. The final design uses a
more expensive composite dielectric consisting of polytetraflouroethylene (teflon) and
ceramic, with an εr of 10.2. W1 and W2 are both 2 mm in width and the separation is
0.12 mm. The calculated characteristic impedance is 49.35 Ω.
The measurement of an impedance using the network analyser requires the establish-
ment of a reference plane, and this is normally performed by placing a short-circuit at
the end of the measurement cable, with all parts of the measurement cable up to that
point being of uniform 50 Ω characteristic impedance.
However for some of the measurements undertaken, there is some non-uniformity of
structure between the end of the coaxial cable and the pair of antenna terminals at
Page 265
9.5 Discussion and Improvement
which we would like to establish a reference plane. If during calibration we place a
shorting pin across the antenna terminals will have, because of the small size of the
pin, not a complete short-circuit but some small inductance.
To remove the effect of this inductance from the measurement, we employ the nor-
malising function of the network analyser that has the effect of making this residual
inductance appear as a perfect short-circuit. Then, without further normalising of the
analyser, the short-circuit is removed and the measurement is taken.
These steps are implemented with the use of a special shorting connection, i.e. some-
thing that resembles the AUT but with a shorting pin across the positive and negative
terminals.
This method is considered to be a practical but approximate method of dealing with the
problems of not being able to maintain a uniform 50 Ω characteristic impedance in a
measurement cable up to the antenna terminals and not being able to place a complete
short-circuit across those terminals.
The results obtained using a Balun are summarised in Table 9.1. It can be seen that
with the use of Balun, the measurement of resistance is the best, though not much
better, than the rest. On the other hand, a high quality co-axial cable provides the best
results for reactance measurement.
Table 9.1. Average error and maximum error. Eave is the average error; Emax is the maximum
error.
Eave (Ω) Emax (Ω)
Re Im Re Im
Cable (1) 2.55 3.38 3.72 6.49
Cable (2) 3.29 3.64 3.96 8.03
Cable (3) 4.91 0.59 6.38 1.62
With Ground Plane 2.62 1.63 4.04 2.91
With Balun 2.35 4.12 3.90 5.26
It is inconclusive as to which measurement method is the best among all, as a direct
measurement using a normal coaxial cable offers comparably acceptable results. As
Page 266
Chapter 9 Measurement Techniques
AUT
RF Transformer
NetworkAnalyser
(a)
S
W
t
W
h
1 2
(b)
(c)
Figure 9.8. The designed Balun. The use of Balun to connect a balanced load to unbalanced
coaxial cable: (a) Top view of the design. (b) The design parameters of coplanar strips.
(c) A picture where the designed Balun is connected to a AUT.
can be seen from Fig. 9.6 and Fig. 9.7, the worst results are still within 10 Ω for the both
resistance and reactance measurements within 860 - 960 MHz band.
Note that the results are only applicable for small RFID antennas. Larger antennas,
such as a half-wave dipole, will most certainly require a Balun or ferrite sleeve for
proper measurement.
Again, in all the comparisons, simulated results are used as the standard values. It is
assumed that with proper fabrication process, an actual antenna with similar charac-
teristics with the modelled antenna, can be obtained. In short:
1. Measurement accuracy drops as frequency increases.
2. From the results shown, a ferrite sleeve does improve the measurement results.
However, it is discovered that the placement of the ferrite sleeve especially the
distance away the ferrite sleeve from the antenna under test, may affect the mea-
surement results.
Page 267
9.6 Effect of Inaccuracy in Measurement
3. Half of the AUT on ground plane gives best improvement overall. However, only
some antennas can be cut into half, such as those that are symmetrical at the feed
point.
4. A shorted antenna, rather then a normal shorted SMA connector, is used for short
circuit calibration of a network analyser before any measurement is carried out.
The investigation of RF cable effects on RFID tag antenna impedance measurement
exposes the difficulties in getting accurate measurement and obtains the best match
between AUT and chip. In actual RFID tag design, it is best to:
1. Design by simulation. Adjust the dimensions of the antenna so that the input
impedance equals to the conjugate of the chip impedance.
2. Fabricate the designed antenna. Measure the antenna input impedance.
3. Attach the tag chip on the antenna. Fine-tune the antenna, such as by reducing
the dimension of certain part of the antenna, while monitoring the read range of
the tag.
9.6 Effect of Inaccuracy in Measurement
In the previous section, it was inconclusive as to which measuring method is the best.
However, with the proper experimental setup, the measurement error when compared
with the simulated results is within a small margin. In this section, the effect of mis-
match caused by the inaccuracy in measurement of the antenna impedance is pre-
sented. The maximum power transfer equation is used to examine this effect, and
the results are presented in graphical form. The analysis presented can also be applied
to any degree of mismatch, as it is not limited to just a small mismatch.
Maximum power transfer is introduced in Section 6.2. When there is no conjugate
match between the RFID antenna and the chip, the power loss can be computed through:
PlostPavailable
=∣∣∣∣Zant − Z∗
cctZant + Zcct
∣∣∣∣2
(9.1)
Page 268
Chapter 9 Measurement Techniques
where Zcct is the impedance of chip and Zant is the impedance of AUT. When Zcct
equals Zant, (9.1) will be zero, or in other words, all the available power is transferred
from the antenna to the chip and there is no power loss.
Both the impedances, Zcct and Zant can be expressed in terms of resistance and reac-
tance:
Zcct = a + jb (9.2)
Zant = c + jd (9.3)
If the reactance of the AUT and the chip are equal and opposite (b = −d), (9.1) can be
simplified to:
PlostPavailable
=∣∣∣∣ a − ca + c
∣∣∣∣2
(9.4)
If resistance of the AUT and the chip are equal (a = c), (9.1) becomes:
PlostPavailable
=∣∣∣∣ (b + d)2 + j2a(b + d)
4a2 + (b + d)2
∣∣∣∣2
(9.5)
or in magnitude form:
PlostPavailable
=(b + d)4 + 4a2(b + d)2
[4a2 + (b + d)2]2(9.6)
Note that b is always negative as an RFID tag chip is always capacitive in nature due to
the reservoir capacitor located in the rectifier circuit. Hence the term (b + d) is actually
the difference between the magnitudes of Im(Zcct) and Im(Zant). This term should be
minimised for maximum power transfer as maximum power transfer occurs when b
equals −d (when (b + d) equals zero), or in other words Zcct is the conjugate of Zant.
The power transfer efficiency can be computed through:
Page 269
9.6 Effect of Inaccuracy in Measurement
η = 1 − PlostPavailable
(9.7)
Power Transfer Efficiency (Equal Reactance)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-40 -30 -20 -10 0 10 20
Difference in Resistance (Rcct-Rant) (Ohm)
Eff
icie
ncy
20
15
10
5
Figure 9.9. The power transfer efficiency between AUT and chip when reactances are equal
and opposite. This graph is based on (9.4) and (9.7). The chip input resistance is
indicated by the small number besides each traces in the graph.
The plots of (9.7) with PlostPavailable
given as (9.4) and (9.6) are presented as Fig. 9.9 and
Fig. 9.10 respectively. These graphs serve two purposes. Firstly, they show the power
transfer loss when there is a mismatch in either the real part or imaginary part between
the AUT and the chip input impedance. It can also be seen from both graphs (Fig. 9.9
and Fig. 9.10) that the resistance of a UHF chip has great impact on the power trans-
fer efficiency. A UHF chip with a higher resistance value has a better power transfer
efficiency when compared with a UHF chip with a lower resistance value, though the
degree of mismatch is the same for these two cases.
The second purpose is to show the impact when the characteristics of the modelled
antenna are not accurately translated into the fabricated antenna. If the initial antenna
design is entirely carried out using a simulation program, this problem will most likely
happen. It shows that by using a higher resistance value chip, the power transfer loss
can be minimised.
Page 270
Chapter 9 Measurement Techniques
Power Transfer Efficiency (Equal Resistance)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-30 -20 -10 0 10 20 30
Reactance Difference (Xcct-Xant) (Ohm)
Eff
icie
ncy
20
15
10
5
2
Figure 9.10. The power transfer efficiency between AUT and chip when resistances are
equal. This graph is based on (9.6) and (9.7). The chip input resistance is indicated
by the small number besides each traces in the graph.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-30 -20 -10 0 10 20
Difference in Both Resistance and Reactance (Ohm)
Eff
icie
ncy
Estimation
Power Loss Curve
Figure 9.11. The comparison between the estimated and calculated values on the power loss
curve. The estimated values are obtained by multiplying (9.4) and (9.6). Calculated
values are based on (9.1). The resistance of the chip is set to 20 Ω in this case.
Page 271
9.7 Conclusion
In the case where there are mismatches in both resistance and reactance (which is in
fact a common scenario), the power transfer efficiency can be computed using (9.1).
Alternatively it can be approximated, even if not very accurately, by multiplying the
two power transfer loss equations (9.4) and (9.6).
An example is shown as Fig. 9.11. In this example, the resistance of the chip (Rcct) is
chosen to be 20 Ω. The x-axis shows the difference of both resistance (Rcct - Rant) and
reactance (Xcct - Xant). For example, at a point in the graph corresponding to value -10
of the x-axis, it simply means that the difference the resistance and reactance between
a tag chip and a tag antenna are both 10 Ω. Note that this graph does not correspond
to the multiplication of Fig. 9.9 and Fig. 9.10, as those graphs either assumed equal
resistance or equal and opposite reactance.
9.7 Conclusion
Detailed investigation of RF cable effects on UHF RFID tag antenna impedance mea-
surement was presented in this chapter. It offers an insight on the difficulties in ob-
taining the correct measurement of impedance. Measurements were carried out using
several different configurations. It is shown that it is impossible to obtain the exact sim-
ulation results (the ideal case). Nonetheless, the impact of the inaccuracy of impedance
measurement on the maximum power transfer between an antenna and a chip can be
predicted. As long as a tag antenna is designed to be able to fine-tuned easily, the chal-
lenges in measuring an antenna’s impedance should not impede the development and
deployment of an RFID tag antenna.
Page 272
Chapter 10
Thesis Conclusions
CONCLUSIONS of this thesis are presented in this chapter.
It reviews and summarises all the contributions from the re-
search during the candidature period. Also, some possible
future research extensions are discussed. It is hoped that this research will
contribute to the development of RFID technology until the vision of ubiq-
uitous identification is achieved.
Page 273
10.1 Introduction
10.1 Introduction
This chapter presents the conclusions of this thesis. A brief overall review is compiled
as Section 10.2. A few interesting research extensions are discussed in Section 10.3.
The summary of the contributions that arise from the research work published by this
thesis, are listed in Section 10.4, followed by the conclusion of this thesis.
10.2 Review of Thesis
The title of this thesis is “Antenna Positioning Analysis and Dual-Frequency Antenna
Design of High Frequency Ratio for Advanced Electronic Code Responding Labels”.
As explained in the introduction chapter of this thesis (Chapter 1), this thesis consists
of two major parts:
1. Antenna positioning analysis.
2. RFID tag antenna design (including novel dual-frequency antenna).
For the first part (Chapter 3 to Chapter 5), research work was devoted to the study
of path loss models and the development of a simulation software to investigate the
signal strength of RFID readers within a certain area. The developed software was
varied to contribute to numerous areas (Refer Section 10.4).
It is regrettable that actual reader deployment in a warehouse was not able to be carried
out during the research period due to a lack of RFID readers. Nonetheless small-scale
testing (one to two readers) has shown that the model suggested gives good estimation
of reader antenna signal strength.
In the second part (Chapter 6 to Chapter 9), designs of HF and UHF RFID antennas
for various new industrially driven applications are described. This is followed by
the design and testing of novel dual-frequency RFID tag antenna, which operates at
both HF and UHF. Also included are the measuring techniques used when measuring
antenna impedances during the prototyping stage.
Page 274
Chapter 10 Thesis Conclusions
The research work on tag antenna designs covers the theoretical tag antenna design,
tag antenna simulation, tag antenna prototyping and measurement, the manufacturing
of fully functional RFID tags and on-site testing. Optimisation of antenna size and
read range performance was carried out when possible, but is not the priority of the
research, as the main aim in those sections was to show feasibility and flexibility in the
designing of tag antennas for hostile environments.
Also it is unavoidable that when comparing free space read range (not in hostile en-
vironment), the RFID tags with novel antennas designed and presented in this thesis
are outperformed by commercial tags. Nonetheless, normal commercial tags would
not be operational in a hostile environment while the newly designed tags can offer an
acceptable read range performance.
Overall, this thesis provides an in-depth investigation of collision problems in dense
reader environments, the tag attachment problem in hostile environments, and the
problem of development of a compact dual-frequency antenna of high frequency ratio.
This work hopefully will contribute to the engineering community which is working
towards an ubiquitous identification system in supply chains.
10.3 Possible Research Extensions
The contents presented in this thesis covers the work carried out within the three-year
candidature time span. Though all the research aims have been achieved with results
and findings published in international level conferences and journals, there are still
some interesting topics for further study. Not really related to the studies of antenna
design and electromagnetic, and in no particular order of importance they are:
1. Real-time Simulator
The core of the simulation software, which compute the signal strength of reader
antennas has been developed and presented in this thesis. However, the develop-
ment of a real-time simulator can be beneficial to real-time adjustment of antenna
placement in a dense RFID environment.
Page 275
10.4 Summary of Contributions
2. Dual-frequency RFID Protocol
The proposed dual-frequency RFID tag antenna will work in HF (using ISO/IEC
18000-3:2004 [33]) and UHF (EPCglobal C1G2 [35]). This research extension is to
examine both the above stated protocols and to develop a combined protocol.
3. Dual-frequency Chip Design
Prototypes of dual-frequency antennas have been proven functioning at both the
HF and UHF bands. However, there is no available dual-frequency chip operat-
ing at these two bands. The design of a dual-frequency chip will require exper-
tise in semiconductor design and RFID design, in integrating the HF standards
(ISO/IEC 18000-3:2004) [33], and the UHF standards (EPCglobal C1G2) [35] to
produce a single chip with a single feed point.
4. RFID Cost Reduction
This a research area in manufacturing engineering, in which new manufacturing
techniques are to be developed based on creative choice of materials.
10.4 Summary of Contributions
1. Antenna Positioning Analysis
A comprehensive study of the RFID collision problems (which focus on the Reader
Interference with Tag Replies (RITR) problem and the Regulatory Reader Shut-
down (RRS) problem), and the development of a path loss model to predict signal
strength radiated by RFID reader antennas was carried out. A piece-wise linear
path loss model is suggested for simulation purposes.
2. Development of Simulation Software
Software written in MATLAB is presented. This software simulates the signal
strength of RFID reader antennas in a defined zone. The simulated results can
be used for reader antenna placement to minimise interference between reader
antennas, maximise coverage, while adhering to local regulations.
Page 276
Chapter 10 Thesis Conclusions
3. Investigation of Sources of Simulation Errors
Intensive research was carried out to identify the common sources of simula-
tion errors, especially in the case of RFID reader antenna simulation in a defined
vicinity. Guidelines were presented to help anyone who is interested in develop-
ing simulation software for RFID deployment.
4. Reader Synchronisation
The idea of reader synchronisation is examined. Reader synchronisation is in-
troduced to allow better RFID performance and coverage when strict regulations
are imposed on RFID deployment in European countries. Various methods were
discussed with several new fine-tuning techniques presented.
5. Regulation and Human Safety
Work was carried out to simulate the radiation power density in the surround-
ings of a dense RFID reader environment. This research was primarily a consul-
tancy on Specified Absorption Rate (SAR). The report generated was essential to
support the experimental work carried out for the approval of 4 W EIRP Radio
Frequency Identification (RFID) license in Australia.
6. Threshold of Second Carrier Sensing
This is an extension study of an existing newly proposed protocol to reduce the
Reader Induced Tag Confusion (RITC) problem in a dense reader environment.
Based on simulation results, recommendations were made on the threshold val-
ues for second carrier sensing with various antenna configurations and place-
ments.
7. HF and UHF RFID Tag and System Design
Pallet level HF and UHF RFID tag antennas are readily available. However, there
are still huge challenges in item level tagging for many types of consumer items,
such as items with ionised liquid (water) or with metallic parts. Throughout
the research candidature, several case studies were completed with the designs,
prototyping and testing of industrially driven RFID HF and UHF tag antennas.
Page 277
10.5 Conclusion
8. Investigation on Measuring Techniques
Several measuring techniques were designed and tested to obtain the impedance
of prototype HF and UHF RFID antennas. This is a well-known problem to RFID
tag antenna designer to measure the impedance of a prototype antenna.
9. Deployment of RFID Tags in Piggery
HF and UHF RFID tags were deployed in a piggery for the first time in Australia.
This field trial demonstrates that the use of RFID technology is beneficial to the
pig industry, as when the eating behaviour of pigs is monitored, the health of
pigs can be inferred from the eating frequencies.
10. Novel Dual-Frequency RFID Tag Antenna
An RFID tag antenna which operates in both HF and UHF RFID bands was de-
signed. This first of its kind antenna has a single feed point and is ready for
wide-scale adoption when a dual-frequency RFID chip is commercially avail-
able. Being a dual-frequency antenna, it embraces both the benefits offered by
HF and UHF antennas. For completeness, several variants of this dual-frequency
antenna were explored, including one with a multi-feed point.
10.5 Conclusion
This chapter identifies all the research work presented in this thesis, which contributes
to RFID practice in recommending various reader antenna positioning and deploy-
ment techniques, design and analysis of HF and UHF tag antennas for hostile environ-
ments, and development of a novel dual-frequency RFID tag antenna.
Page 278
Appendix A
Calculation of Impedanceof Coplanar Strip Line
The characteristic impedance of a coplanar strips (CPS) is computed with reference
to [136].
The variables defined in Fig. 9.8(b) are S, W1, W2, t, h.
To simplify the mathematical expression in calculating the characteristic impedance,
two new variables are defined:
a =S2
(A.1)
b =W1 + W2 + S
2(A.2)
Characteristic impedance, Z0:
Z0 =120π√
εe f f
K(k′)K(k)
(A.3)
where:
εe f f = 1 +12(εr − 1)
K(k)K(k′)
K(k′h)K(kh)
(A.4)
k =
√1 −
( ab
)2(A.5)
k′ =√
1 − k2 (A.6)
Page 279
kh =
√1 − sinh2(πa
2h )sinh2(πb
2h )(A.7)
k′h =√
1 − k2h (A.8)
where K is the complete elliptical integral of the first kind.
Page 280
Appendix B
Some Notes on theConcept of Inductance
This appendix records in-sight obtained from discussion with Prof. Peter Cole on the
concept of inductance. It complements the material of Section 6.10.2.
1. Objectives
The objective of these notes is to clarify some concepts concerned with mutual
and self inductance. This appendix aims to give proper meaning to statements
about inductance that are commonly made without adequate clarification of what
they mean.
2. Some Basics
Inductance, either self or mutual, is defined as flux linkage per unit current. This
definition implies two closed contours: contour C1 for the path of the current,
and contour C2 to define the boundary of the flux linked. We digress to provide
a reminder that the Biot-Savart law commonly quoted for the field of a current
element seems to imply that a current element can act alone to create a field.
However, that is not so. The validity of that law is limited to the case when it
is used to calculate a field by integrating over the full path of the current, which
will flow in a closed contour which we have chosen above to label as C1. But in
this analysis we will not have to use the Biot-Savart law. We will use instead an
expression for magnetic vector potential from which the magnetic flux density
may be derived. But we believe that expression for magnetic vector potential has
its own validity restricted to the case when it is integrated over the full closed
contour of the causing current.
Page 281
3. Calculation of flux linked
The flux Φ2 linked by a contour C2 enclosing a surface S2 through which flows a
flux density B2 derivable from a vector potential A2 is given by:
Φ2 =∫
S2
B2 · ds2 =∫
S2
(curlA2) · ds2 =∮
C2
A2 · dr2 (B.1)
where we have used Stokes’ law to convert from a surface integral to a contour
integral.
Now we can calculate the magnetic vector potential A2(r2) caused by the current
flowing in contour C1 by
A2(r2) =µ0
4π
∮C1
I(r1)dr1
r12(B.2)
where r12 = |r2 − r1| is the scalar distance from the current element I1(r1) to the
potential point r2. Combining the above results and dividing by the current I1
we have the expression for the mutual inductance between the circuit of contour
C1 and the circuit of contour C2. This is
L21 =µ0
4π
∮C2
[∮C1
dr1
r12·]
dr2 (B.3)
Since the limits of the two integrals are independent, we may write this as:
L21 =µ0
4π
∮C2
∮C1
dr1 · dr2
r12(B.4)
We see that L12 = L21, and we call of this result the mutual inductance M. The
formula is known as Neumann’s formula for inductance.
4. Some mathematical difficulties
For contours that do not intersect, there are none. But when contours intersect we
have a difficulty at points where r12 = 0. Furthermore we can not in this simple
form use the result to calculate the self inductance of a circuit.
Page 282
Appendix B Some Notes on the Concept of Inductance
5. Resolution
To deal with these problems, we will reduce the level of idealisation of the prob-
lem by allowing for the non-zero dimensions of that wire carrying the current.
We illustrate in Fig. B.1 the case of a round wire of radius a carrying the current I1,
and consider the radio frequency case when the current will flow on the surface
of the wire. In the diagram we express position vectors r1 and r2 in relation to an
arbitrarily positioned origin.
r2
2a
B r( )2
Arbitrary originO
r1
Figure B.1. Illustration to explain the calculation of flux density of a round wire of radius a.
In calculating at points r2 the flux density B2(r2) and magnetic vector potential
A2(r2) we will make the approximation that the current flows not on the surface
of the wire but at the centre defined in the diagram by the position vector r1. In
calculating the flux linked, we will not let position vector r2 go inside the wire,
but will restrict it to be no nearer the centre than the radius a. Contour C1 now
follows the centre of the wire, and contour C2 now follows the inner edge. Thus
the scalar difference r12 = |r2 − r1| will be no smaller than a, and the integral can
be performed. There are formulae in the literature, for the “self inductance” of
round wires and of wires of various shapes, that are based on this principle.
6. What we find in the literature
Page 283
Among what we can find in the literature are expressions for the self inductance
per unit length of a straight wire, of the given cross sectional shape, or straight
round wire of the given radius.
What this really means is that it (the inductance per unit length) is the contribu-
tion to the integral:
L21 =µ0
4π
∮C2
[∮C1
dr1
r12·]
dr2 (B.5)
which is obtained by integrating (twice) over a length LW of straight wire, mod-
ifying the contours C1 and C2 as described above to ensure that r12 is not zero,
and dividing by the length LW . We find that as LW becomes large, the integral
over the LW part of C1 becomes independent of LW , whereas the integral over
part of C2 is proportional to LW , so a length independent result is obtained after
dividing by LW . But any statement about the “inductance” of the wire segment
must always be understood as being only a contribution to the true inductance
(self or mutual) of a complete circuit (or circuits) and that the complete induc-
tance will consist of the sum of the so-called “inductances” of all wire segments
and so-called “mutual inductances” between all segments.
Page 284
Appendix C
MATLAB Code for PathLoss Calculation
This appendix includes all the MATLAB code required to produce the simulation re-
sults shown in Chapter 3. Detailed explanation on how the code works is also included
in Chapter 3.
Section C.1 shows the main body of the simulation code, where the settings for any
output plot is located. Also, the area of simulation can be changed there. Theoretically,
there is no limit to the size of the area, except the limitation of the computer used for
simulation.
Section C.2 has a data grid to represent the antenna gain pattern for the reader antennas
used in the simulation. If a different antenna is used, a new data grid is required.
The data grid is obtained through actual antenna gain measurement in an anechoic
chamber.
Section C.3 computes the path loss based on path loss model discussed in Chapter 3.
C.1 Main Code
1 c l e a r ;
warning o f f MATLAB: divideByZero ;
3
antenna1 = antgainv2 ( 5 0 0 , 4 5 0 , 5 0 ) ;
5 antenna2 = antgainv2 ( 5 0 0 , 4 0 0 , 5 0 ) ;
t o t a l = 10∗ log10 ( 1 0 . ˆ ( 0 . 1 ∗ ( antenna1 ) ) + 1 0 . ˆ ( 0 . 1 ∗ ( antenna2 ) ) ) ;
7
Page 285
C.2 Antenna Gain Pattern
f igure ( 1 ) =pcolor ( t o t a l ) ;
9 hold ;
11 co lor1 = [0 0 0 . 5 ] ; c o lor2 = [0 0 1 ] ;
co lor3 = [0 1 1 ] ; c o lor4 = [1 1 0 ] ;
13 c o l o r 5 = [1 0 0 ] ; c o l o r 6 = [ 0 . 5 0 0 ] ;
15 colormap ( usercolormap ( color1 , color2 , color3 , color4 , color5 , c o l or 6 ) ) ;
17 c a x i s ([−80 0 ] ) ;
colorbar ;
19 s e t ( gca , ’ PlotBoxAspectRatioMode ’ , ’ manual ’ ) ;
s e t ( f igure ( 1 ) , ’ EdgeColor ’ , ’ none ’ ) ;
21 c l a b e l ( contour ( t o t a l , 2 0 , ’ k ’ ) , ’ manual ’ ) ;
hold ;
C.2 Antenna Gain Pattern
function x=antgainv2 ( size , x , y ) disp ( ’ Loading antenna gain pat te rn &
2 path l o s s . . . . ’ ) ;
n=s ize ; %a r r a y s i z e
4
a=y ; %antenna l o c a t i o n − row
6 b=x ; %antenna l o c a t i o n − column
%antenna f a c i n g down
8
gain=ones ( n , n ) ;
10
%d a t a g r i d
12
ddata=[−14 −13.8500 −13.7000 −13.5500 −13.4000 −13.2500
14 −13.1000 −12.9500 −12.8000 −12.6500 −12.5 −12.3500 −12.2000
−12.0500 −11.9000 −11.7500 −11.6000 −11.4500 −11.3000 −11.1500
16 −11 −10.8000 −10.6000 −10.4000 −10.2000 −10.0000 −9.8000
−9.6000 −9.4000 −9.2000 −9 −8.8500 −8.7000 −8.5500
18 −8.4000 −8.2500 −8.1000 −7.9500 −7.8000 −7.6500 −7.5
−7.3500 −7.2000 −7.0500 −6.9000 −6.7500 −6.6000 −6.4500
20 −6.3000 −6.1500 −6 −5.8500 −5.7000 −5.5500 −5.4000
Page 286
Appendix C MATLAB Code for Path Loss Calculation
−5.2500 −5.1000 −4.9500 −4.8000 −4.6500 −4.5 −4.3500
22 −4.2000 −4.0500 −3.9000 −3.7500 −3.6000 −3.4500 −3.3000
−3.1500 −3 −2.9000 −2.8000 −2.7000 −2.6000 −2.5000
24 −2.4000 −2.3000 −2.2000 −2.1000 −2 −1.8500 −1.7000
−1.5500 −1.4000 −1.2500 −1.1000 −0.9500 −0.8000 −0.6500
26 −0.5 −0.3500 −0.2000 −0.0500 0 .1000 0 .2500 0 .4000
0 .5500 0 .7000 0 .8500 1 1 . 1 1 . 2 1 . 3 1 . 4 1 . 5 1 . 6 1 . 7 1 . 8 1 . 9
28 2 2 . 1 2 . 2 2 . 3 2 .400 2 .500 2 .6000 2 .700 2 .8000
2 .900 3 3 .100 3 .2000 3 .300 3 .4000 3 .500 3 .6000
30 3 .700 3 .8000 3 .900 4 4 .0500 4 .1000 4 .1500 4 .2000
4 .2500 4 .3000 4 .3500 4 .4000 4 .4500 4 . 5 4 .5500
32 4 .6000 4 .6500 4 .7000 4 .7500 4 .8000 4 .85000 4 . 9 0
4 . 9 5 5 5 .0500 5 .1000 5 .1500 5 .2000 5 .2500 5 .3000
34 5 .3500 5 .4000 5 .4500 5 . 5 5 .5500 5 .6000 5 .6500
5 .7000 5 .7500 5 .8000 5 .8500 5 .9000 5 .9500 6 6 6 6
36 6 6 6 6 6 6 6 ] ;
38 ddata back =[−14 −14.1000 −14.2000 −14.3000 −14.4000 −14.5000
−14.6000 −14.7000 −14.8000 −14.9000 −15 −15.1000 −15.2000
40 −15.3000 −15.4000 −15.5000 −15.6000 −15.7000 −15.8000 −15.9000
−16 −16.2000 −16.4000 −16.6000 −16.8000 −17.0000 −17.2000
42 −17.4000 −17.6000 −17.8000 −18 −18.2000 −18.4000 −18.6000
−18.8000 −19.0000 −19.2000 −19.4000 −19.6000 −19.8000 −20
44 −20.0500 −20.1000 −20.1500 −20.2000 −20.2500 −20.3000 −20.3500
−20.4000 −20.4500 −20.5 −20.6500 −20.8000 −20.9500 −21.1000
46 −21.2500 −21.4000 −21.5500 −21.7000 −21.8500 −22 −22.2000
−22.4000 −22.6000 −22.8000 −23.0000 −23.2000 −23.4000 −23.6000
48 −23.8000 −24 −23.9000 −23.8000 −23.7000 −23.6000 −23.5000
−23.4000 −23.3000 −23.2000 −23.1000 −23 −22.9500 −22.9000
50 −22.8500 −22.8000 −22.7500 −22.7000 −22.6500 −22.6000 −22.5500
−22.5 −22.4500 −22.4000 −22.3500 −22.3000 −22.2500 −22.2000
52 −22.1500 −22.1000 −22.0500 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22
−22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22
54 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22
−22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22
56 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22
−22 −22 −22];
Page 287
C.3 Path Loss
58
%c o m p u t a t i o n
60 for i =1 :n
for j =1 :n
62 r a t i o = abs ( ( j−b ) /( i−a ) ) ;
r a t i o = 2∗(90− atan ( r a t i o ) ∗180/ pi ) ;
64 r a t i o = round ( r a t i o ) ;
i f i−a ˜=0
66 i f a<=i
gain ( i , j ) = ddata ( 1 , r a t i o +1) ;
68 else
gain ( i , j ) = ddata back ( 1 , r a t i o +1) ; ;
70 end
else
72 gain ( i , j ) = ddata ( 1 , 1 ) ;
end
74 i f j−b==0&&a<=i
gain ( i , j ) = 6 ;
76 end
end
78 end
x=gain−path loss ( n , b , a , 0 . 0 1 ) ; %r e t u r n ga in
80 disp ( ’Done ’ ) ;
C.3 Path Loss
1 function x=path loss ( size , x , y , u n i t s )
3 n=s ize ; %a r r a y s i z e
ploss = ones ( n ) ;
5 d i s t a n c e = ones ( n ) ;
a=y ; %antenna l o c a t i o n − row
7 b=x ; %antenna l o c a t i o n − column
9 %G e n e r a t e D i s t a n c e
for i = 1 : n
11 for j = 1 : n
d i s t a n c e ( i , j ) =( sqr t ( ( a−i ) . ˆ 2 + ( j−b ) . ˆ 2 ) ) ;
Page 288
Appendix C MATLAB Code for Path Loss Calculation
13 d i s t a n c e ( i , j ) = d i s t a n c e ( i , j ) .∗ u n i t s ;
end
15 end
17 for i = 1 : n
for j = 1 : n
19
i f d i s t a n c e ( i , j ) ==0
21 d i s t a n c e ( i , j ) =1.∗ u n i t s ;
end
23
25 i f d i s t a n c e ( i , j )<8
ploss ( i , j ) =32+25∗ log10 ( d i s t a n c e ( i , j ) ) ;
27 else
ploss ( i , j ) =23+35∗ log10 ( d i s t a n c e ( i , j ) ) ;
29 end
end
31 end
33 x = ploss ;
Page 289
Page 290
Appendix D
MATLAB Code for HFSSVB Script Generation
This appendix shows the MATLAB code written to generate HF coils of different size
to be simulated in Ansoft HFSS. The code based on the API produced by [107].
1 function makecoil ( turn , L , L1 , th ick , height ,w, g , s a v e l o c )
3 % add p a t h s t o t h e r e q u i r e d m− f i l e s .
addpath C: \ Matlab\ h f s s a p i \3dmodeler ;
5 addpath C: \ Matlab\ h f s s a p i \general ;
addpath C: \ Matlab\ h f s s a p i \ a n a l y s i s ;
7 addpath C: \ Matlab\ h f s s a p i \boundary ;
9 % Temporary F i l e s . Thes e f i l e s can be d e l e t e d a f t e r t h e o p t i m i z a t i o n i s
c o m p l e t e . We have t o s p e c i f y t h e c o m p l e t e pa th f o r a l l o f them .
tmpPr jF i le = ’C:\ Matlab\ h f s s a p i \temp . h f s s ’ ;
11 tmpDataFile = s t r c a t ( ’C:\ Matlab\ h f s s a p i \ ’ , s a v e l o c ) ;
t m p S c r i p t F i l e = ’C: \ Matlab\ h f s s a p i \temp . vbs ’ ;
13
% HFSS E x e c u t a b l e Path .
15 hfssExePath = ’C: \ Progra ˜1\ Ansoft\HFSS9\ h f s s . exe ’ ;
17 % C r e a t e a new temporary HFSS s c r i p t f i l e .
f i d = fopen ( tmpScr iptF i le , ’wt ’ ) ;
19
% C r e a t e a new HFSS P r o j e c t and i n s e r t a new d e s i g n .
21 hfssNewProject ( f i d ) ;
h f s s I n s e r t D e s i g n ( f id , ’ t e s t b o x ’ ) ;
23
Page 291
% C r e a t e t h e C o i l .
25
% v a r i a b l e s :
27 n = turn ;
length0 = L ;
29 length1 = L1 ;
t = t h i c k ;
31 h = height ;
width = w;
33 gap = g ;
35 % c r e a t e s u b s t r a t e
sub x = 2∗ length1 +2∗(n+2) ∗ ( gap+width ) ;
37 sub y = 2∗ length0 +2∗(n+2) ∗ ( gap+width ) ;
39 hfssBox ( f id , ’ Sub ’ ,[− sub x/2,−sub y / 2 , 0 ] , [ sub x , sub y ,−h ] , ’mm’ , ’ FR4 epoxy ’ , ’
t rue ’ ) ;
% f u n c t i o n h f s s C y l i n d e r ( f i d , Name , Axis , Center , Radius , Height , Units )
41 h fssCyl inder ( f id , ’ Via 1 ’ , ’Z ’ , [ width /2 , length0+width /2 , 0 ] , width /3 , −h ,
’mm’ , ’vacuum ’ , ’ t rue ’ ) ;
hfssCyl inder ( f id , ’ Via 2 ’ , ’Z ’ , [ width /2 , length0+width/2+n∗ ( width+gap ) ,
0 ] , width /3 , −h , ’mm’ , ’vacuum ’ , ’ t rue ’ ) ;
43 h f s s S u b t r a c t ( f id , ’ Sub ’ , ’ Via 1 ’ , ’ Via 2 ’ ) ;
45 % c r e a t e c o i l
Count=n∗4+1;
47
% Setup
49 S t a r t L o c =[0 , length0 , 0 ] ; Curr Loc= S t a r t L o c ; Ct =1;
51 while ( t rue ) Arm = s t r c a t ( ’Arm ’ , i n t 2 s t r ( Ct ) ) ;
53 i f Ct==1
%1 s t Arm
55 temp length=length1 ;
temp width=width ;
57 hfssBox ( f id ,Arm, Curr Loc , [ length1 , width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;
Page 292
Appendix D MATLAB Code for HFSS VB Script Generation
e l s e i f Ct>=Count
59 %f i n a l arm
temp length =(n+1)∗width+length1 +(n−1)∗gap ;
61 temp width=width ;
hfssBox ( f id ,Arm, Curr Loc , [ temp length , temp width , t ] , ’mm’ , ’ copper ’ , ’
f a l s e ’ ) ;
63 break ; else
%t o p arm
65 temp length = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;
temp width=width ;
67 hfssBox ( f id ,Arm, Curr Loc , [ temp length , temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;
end
69
Ct=Ct +1; Curr Loc=Curr Loc +[ temp length , temp width , 0 ] ;
71 Arm = s t r c a t ( ’Arm ’ , i n t 2 s t r ( Ct ) ) ;
73 %r i g h t arm
temp length=width ; temp width =( Ct /2)∗width+2∗ length0 +( Ct/2−1)∗gap ;
75 hfssBox ( f id ,Arm, Curr Loc , [ temp length ,− temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ )
;
77 Ct=Ct +1;
Curr Loc=Curr Loc +[ temp length ,− temp width , 0 ] ;
79 Arm = s t r c a t ( ’Arm ’ , i n t 2 s t r ( Ct ) ) ;
81 %down arm
temp length = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;
83 temp width=width ;
hfssBox ( f id ,Arm, Curr Loc ,[− temp length ,− temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’
) ;
85
Ct=Ct +1; Curr Loc=Curr Loc+[− temp length ,− temp width , 0 ] ;
87 Arm = s t r c a t ( ’Arm ’ , i n t 2 s t r ( Ct ) ) ;
89 %l e f t arm
temp length=width ; temp width =( Ct /2)∗width+2∗ length0 +( Ct/2−1)∗gap ;
Page 293
91 hfssBox ( f id ,Arm, Curr Loc ,[− temp length , temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ )
;
Ct=Ct +1;
93 Curr Loc=Curr Loc+[− temp length , temp width , 0 ] ;
end
95
% f u n c t i o n h f s s C y l i n d e r ( f i d , Name , Axis , Center , Radius , Height , Units )
97 h fssCyl inder ( f id , ’ Via Cu 1 ’ , ’Z ’ , [ width /2 , length0+width /2 , 0 ] , width /3 , −h , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;
h fssCyl inder ( f id , ’ Via Cu 2 ’ , ’Z ’ , [ width /2 , length0+width/2+n∗ ( width+gap ) ,
0 ] , width /3 , −h , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;
99
%Underpass
101 hfssBox ( f id , ’ Underpass 1 ’ , [ 0 , length0 ,−h ] , [ width , width ,− t ] , ’mm’ , ’ copper ’ , ’
f a l s e ’ ) ;
hfssBox ( f id , ’ Underpass 2 ’ , [ 0 , length0+gap+width ,−h ] , [ width , n∗width +(n−1)∗gap
,− t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;
103
%Merging Cu
105 temp str= ’ ’ ;
107 for i = 1 : ( n∗4+1)
temp str= s t r c a t ( temp str , ’Arm ’ , i n t 2 s t r ( i ) ) ;
109 i f i ˜= ( n∗4+1)
temp str= s t r c a t ( temp str , ’ , ’ ) ;
111 end end
113 h fssUni te ( f id , ’ Underpass 1 ’ , ’ Underpass 2 ’ , ’ Via Cu 1 ’ , ’ Via Cu 2 ’ ,
temp str ) ;
115 %Lump p o r t
hfssRec tangle ( f id , ’ Rec t 1 ’ , ’Z ’ , [ 0 , length0+width ,−h ] , width , gap , ’mm’ ) ;
117
hfssAssignLumpedPort ( f id , ’ LumpedPort ’ , ’ Rect 1 ’ , [ width /2 , length0+width ,−h
] , [ width /2 , length0+width+gap,−h ] , ’mm’ , 5 0 , 0 ) ;
119
% Add an AirBox .
Page 294
Appendix D MATLAB Code for HFSS VB Script Generation
121 hfssBox ( f id , ’ AirBox ’ , [−1.1∗ sub y , −1.1∗ sub y , −1.1∗ sub y ] , [ 2 ∗ 1 . 1 ∗ sub y
, 2 ∗1 . 1∗ sub y , 2∗1 .1∗ sub y ] , ’mm’ , ’vacuum ’ , ’ t rue ’ ) ;
123 % Add PML Layer
%f p r i n t f ( f i d , ’ oModule . CreatePML Array (” UserDrawnGroup : = ” , f a l s e , \n ’ ) ;
125 %f p r i n t f ( f i d , ’” PMLFaces : = ” , Array ( 1 2 9 9 , 1 3 0 0 , 1 3 0 1 , 1 3 0 2 , 1 3 0 3 , 1 3 0 4 ) , \n ’ ) ;
%f p r i n t f ( f i d , ’” C r e a t e C o r n e r O b j s : = ” , t rue , \n ’ ) ;
127 %f p r i n t f ( f i d , ’” T h i c k n e s s : = ” , ”80mm” , ” RadDist : = ” , ”20mm” , \n ’ ) ;
%f p r i n t f ( f i d , ’” UseFreq : = ” , t rue , ” MinFreq : = ” , ”0 .01356GHz”) \n ’ ) ;
129
% Add a S o l u t i o n Setup .
131 %h f s s I n s e r t S o l u t i o n ( f i d , ’ Setup13 56MHz ’ , 13560000/1 e9 ) ;
%h f s s I n t e r p o l a t i n g S w e e p ( f i d , ’ Sweep10to20MHz ’ , ’ Setup13 56MHz ’ , 10000000/1
e9 , 20000000/1 e9 , 201) ;
133
135 % Save t h e p r o j e c t t o a t emporary f i l e and s o l v e i t .
%h f s s S a v e P r o j e c t ( f i d , t m p P r j F i l e , t r u e ) ;
137 %h f s s S o l v e S e t u p ( f i d , ’ Setup13 56MHz ’ ) ;
139 % Expor t t h e d a t a
%hf s s E x p o r t N e t w o r k D a t a ( f i d , tmpDataFi l e , ’ Setup13 56MHz ’ , ’ Sweep10to20MHz ’ ) ;
141
% C l o s e t h e HFSS S c r i p t F i l e .
143 f c l o s e ( f i d ) ;
145 disp ( ’VBS f i l e ready . ’ ) ;
%h f s s E x e c u t e S c r i p t ( h f s s E x e P a t h , t m p S c r i p t F i l e ) ;
Page 295
Page 296
Appendix E
MATLAB Code forInductance Calculation
This appendix includes all the MATLAB code required to compute the inductance of
planar metal track.
Section E.1 is the core of the program. It will accept parameters which will define the
dimensions of a square loop antenna, including number of turns. It will then con-
struct a square loop antenna using segment of straight planar track. The inductance of
each planar track will then be computed using the function shown in Appendix E.2.
Positive and negative mutual inductances between each track are computed by func-
tions shown in Appendix E.3 and Appendix E.3 respectively. The total inductance is
computed by adding the inductance of each track with the total positive mutual induc-
tances and subtracting the total negative mutual inductances.
Section E.2 calculates the inductance of a track, not including the positive and negative
mutual inductances.
Section E.3 calculates the positive mutual inductance between all the tracks.
Section E.4 calculates the negative mutual inductance between all the tracks.
E.1 Main Code
c l e a r a l l ;
2 %in cm
global temp length ;
4 global d e l t a ;
Page 297
E.1 Main Code
global length ;
6 global length1 ;
8 width = 0 . 1 ;
gap = 0 . 0 2 ;
10 length =1;
length1 =2;
12 t h i c k = 0 . 0 0 1 8 ;
n=1;
14 d e l t a =gap+width ;
16 % c r e a t e c o i l
Count=n∗4+1;
18
% Setup
20 S t a r t L o c =[0 , length , 0 ] ;
Curr Loc= S t a r t L o c ;
22 Ct =1;
24 while ( t rue )
26 i f Ct==1
%1 s t Arm
28 temp length ( Ct ) =length1 ;
d i r e c t i o n ( Ct ) =1;
30 e l s e i f
Ct>=Count
32 %f i n a l arm
temp length ( Ct ) =(n ) ∗width+length1 +(n−1)∗gap ;
34 d i r e c t i o n ( Ct ) =1;
break ;
36 else
%t o p arm
38 temp length ( Ct ) = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;
d i r e c t i o n ( Ct ) =1;
40 end
Page 298
Appendix E MATLAB Code for Inductance Calculation
42 Ct=Ct +1;
44 %r i g h t arm
temp length ( Ct ) =( Ct /2)∗width+2∗ length +( Ct/2−1)∗gap ;
46 d i r e c t i o n ( Ct ) =2;
48 Ct=Ct +1;
50 %down arm
temp length ( Ct ) = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;
52 d i r e c t i o n ( Ct ) =3;
Ct=Ct +1;
54
%l e f t arm
56 temp length ( Ct ) =( Ct /2)∗width+2∗ length +( Ct/2−1)∗gap ;
d i r e c t i o n ( Ct ) =4;
58
Ct=Ct +1;
60
end
62
%c a l c u l a t e t h i n f i l m i n d u c t a n c e L 0
64
for i =1 : Count
66
t h i n f i l m ( i ) =induc ( temp length ( i ) , width , t h i c k ) ;
68
end
70
L=sum( t h i n f i l m ) ;
72 %Compute ( + ) mutual p a i r
temp ct =1;
74
for i =1 : Count−4
76
j = i ;
78
Page 299
E.1 Main Code
while j <=Count−4
80 j = j +4 ;
p pa i r ( temp ct , 1 ) = i ;
82 p pa i r ( temp ct , 2 ) = j ;
temp ct=temp ct +1;
84
end
86 end
88 %Compute (−) mutual p a i r
temp ct =1; i =Count ; while i >0
90
for j =1 : Count ;
92 i f d i r e c t i o n ( j ) ==3
n pa i r ( temp ct , 1 ) = i ;
94 n pa i r ( temp ct , 2 ) = j ;
temp ct=temp ct +1;
96 end
98 end
i =i −4;
100 end
102 i =Count−3;
104 while i >0
106 for j =1 : Count ;
i f d i r e c t i o n ( j ) ==4
108 n pa i r ( temp ct , 1 ) = i ;
n pa i r ( temp ct , 2 ) = j ;
110 temp ct=temp ct +1;
end
112 end
i =i −4;
114 end
Page 300
Appendix E MATLAB Code for Inductance Calculation
116 %compute p p a i r
sum p pair =0;
118
for i =1 : s ize ( p pa i r )
120
arm1=p pair ( i , 1 ) ;
122 arm2=p pair ( i , 2 ) ;
124 sum p pair=sum p pair+mutual pair pos ( arm1 , arm2 ) ;
end
126
128 %compute n p a i r
sum n pair =0;
130
for i =1 : s ize ( n pa i r )
132
arm1=n pai r ( i , 1 ) ;
134 arm2=n pai r ( i , 2 ) ;
sum n pair=sum n pair+mutual pair neg ( arm1 , arm2 ) ;
136 end
138 L ;
2∗ sum n pair ;
140 2∗ sum p pair ;
L t o t a l =L+2∗( sum p pair/1000−sum n pair /1000) ;
E.2 Inductance Calculation
1 %C a l c u l a t e i n d u c t a n c e o f t h i n f i l m i n d u c t o r
%
3 %l e n g t h in cm
%width in cm
5 %t h i c k in cm
%
7
function x = induc ( length , width , t h i c k )
9
Page 301
E.3 Mutual Inductance - Positive
l =length ; w=width ; t = t h i c k ;
11
temp =0.002∗ l ∗ ( log (2∗ l /(w+ t ) ) +0.50049+(w+ t ) /3/ l ) ;
13
x=temp ;
E.3 Mutual Inductance - Positive
%C a l c u l a t e Mutual I n d u c t a n c e f o r a p a i r o f s t r i p s
2 %
%l e n g t h in cm
4 %gmd in cm
%
6
function x = mutual pair pos ( arm1 , arm2 )
8
global temp length ;
10 global d e l t a ;
12 l t h 1 =temp length ( arm1 ) ;
l t h 2 =temp length ( arm2 ) ;
14
%a v e r a g e and compute s i d e d i s t a n c e
16 s i d e d i f f =abs ( l th1−l t h 2 ) /2;
18 %f i n d t h e s h o r t e r l e n g t h
min length=min ( l th1 , l t h 2 ) ;
20
%compute gmd v a l u e
22 gmd temp=abs ( arm1−arm2 ) /4∗d e l t a ;
24 i f s i d e d i f f ==0
x=mutual ( min length , gmd temp ) ;
26 else
x=mutual ( min length+ s i d e d i f f , gmd temp )−mutual ( s i d e d i f f , gmd temp ) ;
28 end
Page 302
Appendix E MATLAB Code for Inductance Calculation
E.4 Mutual Inductance - Negative
%C a l c u l a t e N e g a t i v e Mutual I n d u c t a n c e f o r a p a i r o f s t r i p s
2 %
%l e n g t h in cm
4 %gmd in cm
%
6
function x = mutual pair neg ( arm1 , arm2 )
8
global temp length ;
10 global d e l t a ;
global length ;
12 global length1 ;
14 l t h 1 =temp length ( arm1 ) ;
l t h 2 =temp length ( arm2 ) ;
16
%a v e r a g e and compute s i d e d i s t a n c e
18 s i d e d i f f =abs ( l th1−l t h 2 ) /2;
20 %f i n d t h e s h o r t e r l e n g t h
min length=min ( l th1 , l t h 2 ) ;
22
%compute gmd v a l u e
24 gmd temp=abs ( arm1−arm2 ) /4∗d e l t a ;
26 % adding gap t o gmd
% i f odd
28 i f rem ( arm1 , 2 ) ==1
gmd temp=gmd temp+2.∗ length ;
30 else
gmd temp=gmd temp+2.∗ length1 ;
32 end
34 i f s i d e d i f f ==0
x=mutual ( min length , gmd temp ) ;
36 else
Page 303
E.4 Mutual Inductance - Negative
x=mutual ( min length+ s i d e d i f f , gmd temp )−mutual ( s i d e d i f f , gmd temp ) ;
38 end
Page 304
Appendix F
Path Loss Experiment
F.1 Preliminary Setting Up Procedure
1. Each of the two antennas was mounted on a tripod stand as shown in Fig. F.1.
Both antennas were positioned at the same height and the distance from the
ground to the centre point of the antenna, h, was measured. The value obtained
was h = 1.29 m.
Figure F.1. Antennas on tripods.
2. A signal of 10 dBm at a frequency of 902 MHz from the HP ESG-3000A Signal
Generator was fed into one antenna. While 2 W ERP is equivalent to +33 dBm,
10 dBm was a convenient output power, well within the capability of the Signal
Generator. In addition, while the frequency of operation in Europe is from 865
Page 305
F.2 Experiment Procedure and Results
to 868 MHz, the antennas that we had available were rated for operation from
902 to 928 MHz. The lower 902 MHz was chosen so as to be within the operating
parameters of our antennas and be as close to the European frequency band.
3. The spectrum analyser was set to a range of 900 to 904 MHz. The other antenna
was connected to the HP 8594E Spectrum Analyser with a 10 dB attenuator at the
input to the spectrum analyser.
F.2 Experiment Procedure and Results
F.2.1 Map
Map showing the locations of measurements:
RFID
Lab
Office
Office
Antenna
(Transmitter)To EM
building
Antenna
(Receiver)
RFID
Lab
Office
Office
Antenna
(Transmitter)To EM
building
Antenna
(Receiver)
(a) (b)
Figure F.2. Measurement of signal strength. (a) within a room, and (b) between rooms in
Engineering North building
F.2.2 Signal Strength at Different Distances (Within and Between
Buildings)
1. With the antennas setup as described above, the second antenna (connected to
the spectrum analyser) was moved around various points in the building, with
propagation through walls and typical building materials. The second antenna
was also moved to other nearby buildings. (Please refer to previous for the path
Page 306
Appendix F Path Loss Experiment
RFID
Lab
Office
Office
ENGINEERING NORTH(West Side)
Path taken when
measuring signal strength
Antenna
(Transmitter)
To Maths
building
To EM
building
To Engineering South
Figure F.3. Path taken when measuring signal strength from the west side of Engineering
North building.
RFID
Lab
Office
Office
Antenna
(Transmitter)
Path taken when
measuring signal strength
To EM
building
ENGINEERING NORTH(East Side)
Figure F.4. Path taken when measuring signal strength from the east side of Engineering
North building.
RFID
LabAntenna
(Transmitter)
From
Engineering
North
Path taken when measuring
signal strength
ENGINEERING & MATHS (EM)BUILDING
Figure F.5. Path taken when measuring signal strength from Engineering and Maths (EM)
building.
Page 307
F.2 Experiment Procedure and Results
taken when measuring the signal strength.) For all measurements, the orientation
of Antenna 1 was adjusted so as to point approximately towards the expected
location of Antenna 2, see map in previous section.
2. The strength of the received signal was recorded and the straight-line distance
from the transmitting antenna to the receiving antenna was approximated at each
point.
3. The free space path loss was calculated using the equation, Ploss = 20 log10
(4πd
l
),
where d is the distance between the transmitting and receiving antennas and l
is calculated to be 0.333 m for a frequency of 902 MHz. Considering a signal
of amplitude 10 dBm, antennas of gain 8 dBi each, cable loss of 1.5 dBm (on
both the transmitter and the receiver sides respectively), a 10 dB attenuation from
the attenuator, and using the free space path loss obtained, the expected signal
strength at various points was calculated.
4. The measured signal strengths obtained from various points were then compared
with their respective calculated (expected) signal strength. The results are shown
in Table F.1 below. The graph of measured and calculated signal strengths against
distance, d, was plotted and is as shown below in Fig. F.6.
Comments on results: From the results in Table F.1, it can be observed that the mea-
sured strength of the received signal was stronger on the east side of the Engineering
North building as compared to the west side. The reason a weaker signal strength was
detected on the west side may be due to the reason that an anechoic chamber is located
in between the propagation path of the signal. It can also be observed from the results
that, when measurement was done in the east end of the Engineering and Maths (EM)
building, the strength of the received signal was slightly better compared to the mea-
surement taken in front of the lift in that same building. This is because, in the east
end, there was an open space in the middle of the signal propagation path, and lots of
glass windows, as opposed to brick walls.
Page 308
Appendix F Path Loss Experiment
Table F.1. Measured and calculated signal strengths for various locations. d is the distance
between antennas; Pcal is the calculated received power; Pmea is the measured received
power.
d (m) Pcal (dBm) Pmea (dBm) Location/Comments
8.25 -37 -45 Within room (N203)
4 -31 -35 Within room (N203)
18 -44 -70 Between rooms (N203 - N220b)
36 -40 -75 West of engineering north building. Attenuator
pad removed.
51 -43 -85 Further west of engineering north building.
52 -43 -90 Further west of engineering north building. Fre-
quency span of spectrum analyser reduced to
901.9 to 902.1 MHz.**
67 -45 -96 Further west of engineering north building.
75 -46 -100 Further west of engineering north building.
25 -37 -63 East of engineering north building.
33 -39 -60 Further east of engineering north building.
44 -41 -74 Further east of engineering north building.
28 -37 -78 At north of EM building (3rd Floor, in front of
lift)
37.5 -40 -71 At north of EM building (3rd Floor, east end,
propagation through glass and open area)
28 -37 -90 At north of EM building (4th Floor, in front of
lift, propagation through floor)
37.6 -40 -85 At north of EM building (4th Floor, east end,
propagation through glass, open area and floor)
Page 309
F.2 Experiment Procedure and Results
Comparison Between Calculated and Measured Received
Signal Strength In Building
-31
-37
-44-46-45
-43-43-41
-40-40-40-39-37-37-37
-90
-100
-96
-90
-85
-74
-85
-71
-75
-60
-78
-63
-70
-45
-35
-120
-100
-80
-60
-40
-20
0
0 10 20 30 40 50 60 70 80
Distance of Separation Between Antennas, d (m)
Calculated
Measured
Straight line approximations
Re
ce
ive
dS
ign
al
Str
en
gth
(dB
m)
Figure F.6. Comparison between calculated and measured received signal strength (in build-
ing).
F.2.3 Reflection from Typical Wall and a Conductive Fence
1. The antennas were setup according to the description in Part (A) above. The
settings for the signal generator remained the same, and the spectrum analyser
was set to a frequency range of 901.9 to 902.1 MHz with the 10 dB attenuator at
the input removed.
2. Both antennas were arranged and positioned to face the wall (Fig. F.7) and the
strength of the received signal was measured. It has to be noted that the antennas
should not be placed too close to each other to avoid cross coupling.
3. Step (2) was repeated except that this time, both the antennas were positioned to
face to the same point on the wall (Fig. F.8).
4. Step (2) was again repeated, but this time with the antennas facing a conductive
fence instead of the wall (Fig. F.9).
Page 310
Appendix F Path Loss Experiment
WA
LL
Transmitting
Antenna
Receiving
Antenna
(a) (b)
Figure F.7. Antennas facing directly perpendicular to the wall.
WA
LL
Transmitting
Antenna
Receiving
Antenna
(a) (b)
Figure F.8. Antennas facing with an angle to the wall.
Figure F.9. Antennas on tripods facing a conductive fence.
5. The results are as shown in Table F.2 below.
Page 311
F.2 Experiment Procedure and Results
Table F.2. Measured signal strength for antennas facing a wall or a conductive fence. Ds is
the straight line distance between antennas; Dp is the perpendicular distance from wall
(or fence); Prec,d is the measured received power when antennas are pointing directly to
wall; Prec,0 is the measured received power when antennas are pointing at same point on
wall.
Ds Dp Prec,d Prec,0
Antennas facing wall (1st) 2.5 2.3 -41 -45
Antennas facing wall (2nd) 2.2 1.5 -45 -38
Antennas facing conductive fence 1.77 1.66 -32 —
F.2.4 Propagation Loss Outdoors
1. This part of the experiment was performed outdoor, on the lawn in front of the
Mathematics building.
2. The antennas were setup in a similar way. The settings for the signal generator
remained the same, and the spectrum analyser was set to a frequency range of
901.9 to 902.1 MHz with the 10 dB attenuator at the input removed.
3. The antennas were arranged and positioned to face each other (Fig. F.10) and the
distance of separation between the two antennas, d, was measured. It was found
that d = 11 m. The height of the antennas from the ground, h, was also measured
and it was found to be 1.29 m.
4. The strength of the received signal was measured.
5. The procedures were repeated with both antennas placed on the ground and fac-
ing upwards (with the distance, d, maintained the same) (Fig. F.11).
6. The recorded results are as shown in Table F.3 below.
Page 312
Appendix F Path Loss Experiment
(a) (b)
Figure F.10. Measurement of signal strength in outdoors.
Figure F.11. Antennas on the ground and facing upwards.
F.3 Conclusion
As expected, the measured values of the received signal strength were found to be
quite different from the calculated values, as obtained from the Free Space Loss Equa-
tion. Fortunately, the observed path loss was greater than the path loss calculated in
free space, due to building materials and objects in various rooms.
Our transmitter operated at +10 dBm, whereas RFID transmitters are allowed to oper-
ate at 2 W ERP, or 35 dBm. The path loss required would then be 25 dB greater than
our measured results, i.e.121 dBm (96 + 25). Using Fig. F.6, a distance of at least 110 m
would be required for the separation of two readers, where one is already operating
in a particular sub-band and the other listening before wanting to talk in the same
sub-band.
Page 313
F.3 Conclusion
Table F.3. Calculated and measured signal strength (outdoor) for different antenna orien-
tations. d is the distance between antennas; Prec is calculated received power; Pmea is
the measured received power.
Antenna orientation Distance d (m) Prec (dBm) Pmea (dBm)
Antennas facing directly at each other 11 -29 -39
Antennas on the ground, facing upwards 11 — -71
Page 314
Appendix G
RFID Deployment inPiggery
This appendix presents all the design steps for RFID deployment in a piggery. In the
first part the modelling of a pig feeder in laboratory is shown. In the second part, the
final report on the RFID deployment in piggery is included, which covers the HF and
UHF tags designed, reader installation at piggery, experiment data and analysis, and
conclusion.
G.1 Modelling of Pig Feeder
Raw Material:
1. 2 × Zinc plate (600mm × 1000mm)
2. 1 × Zinc plate (600mm × 400mm)
3. 2 × Wooden piece (400mm × 900mm)
4. Small Zinc plates
5. Bolts and nuts
Note: The dimensions of Zinc plate (Item 1 and 2) are increased by 50 mm on each side,
so that the extra length can be bent and bolted onto.
Construction of Main Body:
Page 315
G.1 Modelling of Pig Feeder
(a)
A
B
CC
DE
F
G
(b)
Figure G.1. A pig feeder used in RFID deployment experiment. (a) Picture of a pig feeder.
(b) Major parts of a pig feeder to be modelled.
a. Raw materials item 1 (2 × Zinc plates) are bent 50 mm each on 4 sides. The
biggest surface is measured 500 mm × 900 mm.
b. Item 1 (2 × Zinc Plates) are bolted at location A and B onto the two wooden
pieces (Item 3). Refer to Fig. G.1(b) and Fig. G.2 for clarity.
Construction of Feeder Part:
a. Raw material item 2 is bent 50 mm each on 4 sides. The biggest surface is mea-
sured 500 mm x 300 mm. It is bolted in location E.
b. A small Zinc plate is bolted in location D. Refer to Fig. G.1(b) and Fig. G.3 for
clarity. The size is a rough approximation as it depends the on the accuracy of
construction of part A, C and E.
c. A small Zinc plate is bolted in location F. Refer to Fig. G.1(b) and Fig. G.3 for
clarity. The dimension measurement is dependent on A, D, and E.
d. At position G of both sides of the feeder (dotted line), place a Zinc strip (width
of 50 mm). Refer to Fig. G.1(b) and Fig. G.3 for clarity.
Page 316
Appendix G RFID Deployment in Piggery
400
900
90
0
50
70
23
A
B
C
Figure G.2. Modelling of a pig feeder (Stage 1).
G.2 Final Report
This section contains the final report sent to industrial partners based on the field trials
carried out a piggery. This report is slightly modified to avoid repetition. The pictures
appear in the original report has been included in the main body of the thesis and will
be referred to.
G.2.1 Executive Summary
This is a report on the design, manufacture and trial testing of RFID tags for possible
future deployment in a piggery. Two types of tags were produced, employing different
RFID technology, one at HF (13.56 MHz) and the other at UHF (in Australia 920 to
926 MHz). Section G.2.2 presents the design, fabrication and encapsulation of the HF
and UHF RFID tags. Section G.2.3 shows the feeder design and the feeder setup in the
piggery. The Experiment setup is examined in Section G.2.4 with experimental results
Page 317
G.2 Final Report
90
300
315
735
100
D E
F
G
Figure G.3. Modelling of a pig feeder (Stage 2).
discussed in Section G.2.5, followed by conclusions and recommendation for future
studies at the end of this document.
G.2.2 Design of RFID Tag Antennas
This section shows the final design of the RFID tag antenna. RFID tags suitable for
attaching to a pig’s ear must be encapsulated in a strong casing before deployment to
protect the tag antennas from being damaged by the pigs. Hence the design of the tag
antenna is restricted by the size of the casing. As the objective of the project was to
compare the performance of a number of RFID technologies, standard commercially
available Low Frequency (LF) technology tags as marketed by Leader Products were
deployed, and compared with High Frequency (HF) technology and Ultra High Fre-
quency (UHF) technology tags, designed by the Auto-ID Lab, Adelaide. The antennas
on the Printed Circuit Board (PCB) are manufactured by Entech Electronic. For small
Page 318
Appendix G RFID Deployment in Piggery
scale testing, RFID tag chips (both HF and UHF from Texas Instruments) were man-
ually attached onto custom designed RFID tag antennas using Z-axis conductive tape
(Electrically Conductive Adhesive Transfer Tape 9703 from 3M).
1. HF RFID Tag Antenna Design
The final design of the HF antenna is shown in Fig. 6.40. The thickness of the sub-
strate is 0.2 mm, and the coil track separation is 0.3 mm. The chip that used was
the Texas Instrument Tag-it HF-I Standard Transponder IC, supplied by Electro-
Com, which is based on the ISO/IEC 15693 and ISO/IEC 18000-3 Mode 1 stan-
dards, and has a nominal input capacitance of 23.5 pF. The RFID reader used for
identifying HF RFID tags is ID ISC.LR2000 from Feig Electronic.
2. UHF RFID Tag Antenna Design
The final design of the UHF antenna is based on the antenna shown in Fig. 6.49(a).
The thickness of the substrate is 0.2 mm. The antenna is designed to have a
matching impedance with respect to the RFID chip used, which is RI-UHF-STRAP-
08 UHF Gen2 Strap from Texas Instruments, supplied by Electro-Com. The RFID
reader used for identifying UHF RFID tags is ID ISC.LRU2000 from Feig Elec-
tronic.
3. Tag Encapsulation
Tags were encapsulated by Leader Products. Encapsulated tags are shown in
Fig. 6.42(b). Read range performances for both the HF and UHF tags before and
after encapsulation were recorded. It is expected that the encapsulation process
will have a relatively bigger impact on the UHF tags when compared to the HF
tags.
4. Performance of HF RFID Tags
Table G.1 shows the read range performance of HF tags before and after encap-
sulation. Also, after encapsulation, tags are positioned on a human hand to sim-
ulate the actual field trial condition, where tags are attached on the ears of pigs.
As expected, the encapsulation did not affect the tag performance by much. The
performance of tags fluctuates slightly as the fine-tuning was carried manually.
Page 319
G.2 Final Report
Table G.1. Performance of HF RFID tag before and after encapsulation.
Tag No. Free Space (m)After Encapsulation
Free Space (m) On Hand (m)
HF01 0.32 0.32 0.33
HF02 0.18 0.27 0.18
HF03 0.34 0.33 0.25
HF04 0.33 0.32 0.33
HF05 0.32 0.28 0.25
HF06 0.34 0.32 0.35
HF07 0.32 0.31 0.33
HF08 0.25 0.23 0.20
HF09 0.32 0.28 0.34
HF10 0.24 0.26 0.28
HF11 0.29 0.30 0.26
HF12 0.31 0.30 0.30
HF13 0.33 0.33 0.34
HF14 0.26 0.26 0.32
HF15 0.32 0.31 0.35
HF16 0.32 0.31 0.32
HF17 0.33 0.33 0.33
HF18 0.31 0.30 0.33
HF19 0.32 0.30 0.33
HF20 0.33 0.31 0.34
HF21 0.34 0.33 0.25
HF22 0.34 0.34 0.34
HF23 0.34 0.33 0.36
HF24 0.32 0.32 0.26
HF25 0.33 0.32 0.34
HF26 0.33 0.32 0.33
5. Performance of UHF RFID Tags
Similar to Table G.1, Table G.2 shows the read range performance of UHF tags
before and after encapsulation. Also, after encapsulation, tags are positioned on
a human hand to simulate the actual field trial condition, where tags are attached
on the ears of pigs.
Page 320
Appendix G RFID Deployment in Piggery
Table G.2. Performance of UHF RFID tag before and after encapsulation.
Tag No. Free Space (m)After Encapsulation
Free Space (m) On Hand (m)
C1 0.60 0.80 0.34
C2 0.49 1.10 0.33
C3 0.45 0.40 0.30
C4 1.04 1.05 0.32
C5 0.46 0.70 0.35
C6 0.62 0.85 0.30
C7 0.36 0.75 0.34
C8 0.40 0.82 0.43
C9 0.89 0.43 0.40
C10 0.49 0.87 0.36
D1 0.68 0.85 0.36
D2 0.49 2.06 0.38
D3 0.41 1.15 0.40
D4 0.37 2.08 0.40
D5 0.85 0.80 0.47
D6 0.43 0.50 0.30
D7 0.65 0.80 0.31
D8 0.50 0.53 0.33
D9 0.60 0.83 0.37
D10 1.00 0.67 0.35
E1 0.55 0.36 0.33
E2 0.34 0.43 0.24
E3 0.86 0.36 0.31
E4 0.54 0.45 0.29
E5 0.36 1.09 0.30
E6 0.65 0.77 0.29
E7 0.59 1.05 0.40
E8 0.58 0.34 0.29
E9 0.49 0.78 0.35
E10 0.21 0.42 0.30
Page 321
G.2 Final Report
The inconsistency in results is caused by several reasons: (i) All tags are fine-
tuned by hand. (ii) All tags are tuned higher than the operating frequency as
experience shows that encapsulation will de-tune the antenna by reducing the
tag’s resonant frequency.
G.2.3 Feeder Design and Experiment Setup
The trials were conducted at The Department of Primary Industries, Victoria at the Pig
Research & Training Centre, 600 Sneydes Rd, Werribee 3030. Fig. G.1 shows a typical
feeder used at that facility. The objective of the trial is to record when a pig, which has
been tagged with an RFID tag, goes to the feeder to eat. Fig. G.4 is the front view of the
planned feeder with RFID reading capability. Fig. G.5 shows the location of antennas
in the protective casing at both sides of a feeder. The slope of a feeder is made of metal.
The antennas positioned at the sides should not intersect with the metallic slope, as
this will degrade the performance of the RFID tag reading.
The planning for HF and UHF readers in plastic protective casings are shown in Fig. G.6
and Fig. G.7. HF system requires an external multiplexer and an additional power
supply. Fig. 6.44 and Fig. 6.45 are photos taken at the test site, with HF and UHF RFID
system setup and reader antennas embedded in thick plastic protective casing besides
the feeders. UMD was contracted for packaging and installation of the readers and
antennas.
G.2.4 Network Setup
Remote access is set up so that the computer (A personal computer running on Pen-
tium 4 with Windows XP) managing the HF and UHF readers can be controlled from
Adelaide through the Internet. An illustration on how the data retriever system works
is as shown in Fig. G.8.
1. Window Scheduler
Triggers sendlog HF.bat and sendlog UHF.bat everyday 0:30 am.
Page 322
Appendix G RFID Deployment in Piggery
feeder
Change to Non-Metalic Sides
UHF Antenna Box
Reader Box
Cables
Figure G.4. Overview arrangement of RFID deployment in piggery.
2. sendlog HF.bat
Compresses the HF data file one day before, packages the data as an email, and
sends the file to smtp.gmail.com through stunnel.
3. sendlog UHF.bat
Compresses the UHF data file one day before, packages the data as an email, and
sends the file to smtp.gmail.com through stunnel.
4. stunnel
stunnel is a software package used to manage the security of data transmission
using Secure Sockets Layer (SSL) protocol.
Window Scheduler is built-in function of Windows XP. sendlog HF.bat and sendlog UHF.bat
are modified version of software supplied by Behnam Jamali. stunnel is a freeware
available on the Internet.
Page 323
G.2 Final Report
Reader Box
(Side View)
Cable
Slope ofthe feeder
Feeder
l
Antenna Box(Grey Zone)
Figure G.5. The position of an antenna in an antenna box besides a pig feeder.
G.2.5 Experiment Results
The target trial pigs were tagged on 24th September, 2007. Testing was carried from
25th to 27th September, 2007. The times the pigs moved to the feeder to eat were
recorded, so that later the data file could be checked for misreads and ghost reads,
where misreads occur when a tag enters interrogation zone but is not picked up by
reader, and ghost reads refer to the situation when unknown tag numbers appeared in
the log file.
The data files, recording pigs at the feeders were monitored remotely after 27th Septem-
ber, 2007. It is not possible to check for misread, but it is possible to examine the eating
behaviour of individual pigs. If a pig looses its appetite drastically, as evidenced by no
reads on the data file, either the tag has failed, or it has been lost, or the pig may need
medical attention.
Page 324
Appendix G RFID Deployment in Piggery
8080
80
340
330
180
40 40
20
100
a
b
c c
d d e
f f
110
640
20
Figure G.6. Encapsulation of a HF reader.
As shown in Fig. 6.46, when facing a pig, the HF tag is on the left (the pig’s right ear)
while the UHF tag is on the right (the pig’s left ear). When a pig is feeding, as shown
in Fig. 6.47, the HF tag is on the right and the UHF tag is on the left when viewed from
the back of a pig.
1. Initial Test Run
Tags of reasonable performance are randomly picked and attached onto the ears
of pigs. The chosen tags are recorded in Table G.3.
The results from 25th to 27th September 2007 were recorded where:
Page 325
G.2 Final Report
100
320
180
40
20
ab
c d
20
e e
80
440
280
20
Figure G.7. Encapsulation of a UHF reader.
Stunnel
WindowsScheduler
Sendlog_HF.bat
Smtp.gmail.com
Backto Adelaide
Sendlog_UHF.bat
Figure G.8. Network setup in a piggery. Data is automatically sent back to the Adelaide labora-
tory.
• Misread: Tag enter interrogation zone but not picked up by reader.
• Ghost Read: Unknown tag number appeared in the log file.
• Read outside of read zone: Tag not in the designated zone but picked up by
the reader.
Page 326
Appendix G RFID Deployment in Piggery
Table G.3. HF and UHF RFID tags attached to pigs.
Pig HF UHF
1 HF17 C1
2 HF07 D3
3 HF20 D8
4 HF14 D7
5 HF22 D1
6 HF19 C6
7 HF16 D10
8 HF15 E7
9 HF04 C4
10 HF06 C10
Date: 25 Sept. 2007
Time: 1125-1145
HF Misread: 0 misread
HF Ghost Read: 0
HF Read outside of read zone: 1
HF Special Notes: Good performance so far. However, pigs are relatively inac-
tive. More observation needed.
UHF Misread: 0 misread
UHF Ghost Read: 0
UHF Read outside of read zone: 0
UHF Special Notes: Good performance so far. However, pigs are relatively inac-
tive. More observation needed.
Date: 25 Sept. 2007
Time: 1347-1427
HF Misread: 2 misread
HF Ghost Read: 0
HF Read outside of read zone: 2
UHF Misread: 0 misread
Page 327
G.2 Final Report
UHF Ghost Read: 1 (426; EPC Gen 2; 00065090C600EE000C06080000; 14:55:09.834
09/25/2007;|1|;)UHF Read outside of read zone: 0
Date: 26 Sept. 2007
Time: 0931-1034
HF Misread: 2 misread
HF Ghost Read: 0
HF Read outside of read zone: 14
HF Special Notes: High Read outside of read zone were caused by active pigs
running near the side of the RHS antenna. Pigs running near the side of the RHS
antenna are manually recorded and coincide with the pig recorded by the reader.
This shows an obstacle can efficiently reduce the Read outside of read zone.
Recommendation: Put obstacles near the side of the feeder to avoid pig getting
into the unintended read zone.
UHF Misread: 0 misread
UHF Ghost Read: 0
UHF Read outside of read zone: 0
Date: 27 Sept. 2007
Time: 1315-1345
HF Misread: 3 misread within 30 minutes.
HF Ghost Read: 0
HF Read outside of read zone: Several occasions when a pig gets very near to the
outer RHS antenna.
HF Special Notes:LHS Antenna not performing well. This is caused by the nearby
metal sheet. Furthermore, tags are tagged on the left hand side of the pig if
viewed from front (Refer Fig. 6.46), making the detection of the pig on the left
even harder.
UHF Misread: 0 misread
Page 328
Appendix G RFID Deployment in Piggery
UHF Ghost Read: 0
UHF Read outside of read zone: 0
UHF Special Notes: At 1331, Pig 0 jumped and landed on pig 7, which is eating
at the feeder. The reader picked up a single read as Pig 0 extended its head into
the read zone.
2. Distance Monitoring
Data was collected remotely to monitor the eating behaviour of tagged pigs. The
total read count fluctuates due to power failures in the computer room at the
test site. The results are as shown in Table G.4, where the numbers of reads for
each tag (both HF and UHF) are shown. A pig feeding at HF feeder will only be
recorded by the HF reader, not by the UHF reader, and vice versa.
A few observations: (1) UHF tag 8 failed on 30 Sept., 7 days after it was tagged
on the pig ear. It was discovered later that this tag has dropped off the pig ear.
(2) It can be seen that both Pig 6 and Pig 8 are relatively inactive. (3) The other
pigs appear to be eating normally.
G.2.6 Conclusions
Based on the initial test run, which spans 25th to 27th September 2007 with manual
observation:
1. The performance of RFID UHF system is excellent, with no misreads. The only
error recorded is a single ghost read on 25th Sept. 2007. The reason for the ghost
read is unknown. According to the reader manufacturer, FEIG, ghost reads are
extremely rare.
2. However, the performance of the RFID HF system is not as good. Misreads oc-
cur in almost every session. Misread always happens on the left hand side of
the feeder. The reason can be (i) The left hand side antenna is near to a UHF
reader antenna. Although arrangement has been made to shift the antenna fur-
ther away, it is suspected a greater distance is required. (ii) The HF tag is attached
Page 329
G.2 Final Report
Table G.4. Results from remote monitoring.
Date Pig 1 2 3 4 5 6 7 8 9 0 Read
27 Sept.HF 601 348 305 330 569 425 280 38 116 206 3218
UHF 325 489 311 478 316 56 336 159 533 628 3631
28 Sept.HF 344 254 221 146 471 100 153 351 124 287 2451
UHF 406 293 189 296 369 59 304 88 414 411 2829
29 Sept.HF 523 96 289 77 616 123 404 55 11 222 2416
UHF 470 528 352 417 449 143 320 92 314 612 3697
30 Sept.HF 215 38 83 57 319 112 114 20 63 70 1091
UHF 139 206 28 130 65 56 56 2 190 87 959
1 Oct.HF 432 351 394 196 641 233 108 122 88 342 2907
UHF 107 245 63 118 99 59 302 0 93 129 1215
2 Oct.HF 336 329 383 159 544 182 276 73 67 237 2586
UHF 276 502 152 398 175 177 482 0 597 393 3152
3 Oct.HF 88 83 64 32 403 53 125 47 32 48 975
UHF 488 382 297 370 255 196 569 0 519 558 3634
4 Oct.HF 457 651 468 171 573 251 424 131 232 882 4240
UHF 605 305 405 450 469 190 796 0 561 740 4521
to the right ear of a pig. When a pig is eating at the feeder, the HF tag will be ap-
proximately 200 mm from the left antenna and 400 mm from the right antenna.
The left antenna should pick up the tag, but due to the degradation as mentioned
in (i), a misread can occur.
3. The tag encapsulation from Leader Product is excellent. So far all tags are still
functioning properly, and not damaged by pigs through constant chewing.
4. Installations of HF and UHF readers by UMD are of good standard.
Based on the distance monitoring:
1. UHF tag 8 stopped giving read data from 30 September onwards. It was discov-
ered later that UHF tag 8 dropped off from the pig’s ear.
Page 330
Appendix G RFID Deployment in Piggery
Recommendation for Future Studies:
While this trial appears to indicate that the UHF system is more reliable and robust,
and for the foreseeable future, UHF tags are likely to be cheaper that HF tags, the HF
system should not be ruled out. In a future trial it is recommended that more fine-
tuning may be necessary, either to provide more space around the feeders, to reduce
the number of misreads and reads outside the zone, or as suggested by Bruce Dumbrell
from Leader Product, the HF antenna can be put on top or under the feeder entry point,
rather than the current position which is at the sides of the feeder entry point.
Page 331
Page 332
Bibliography
∗Please refer to the Conventions Section (page xxi) for descriptions of the citation style used inthis bibliography.
[1] M. Shell. (2007, 11 Jan.) How to Use the IEEEtran LATEXClass. [Online]. Available: http://tug.
ctan.org/tex-archive/macros/latex/contrib/IEEEtran/IEEEtran HOWTO.pdf [12 Nov. 2007].
[2] M. Shell. (2007, 11 Jan.) The IEEEtran BIBTEX style. [Online]. Available: http://www.ctan.org/
tex-archive/macros/latex/contrib/IEEEtran/bibtex/IEEEtran bst HOWTO.pdf [12 Nov. 2007].
[3] Bureau International des Poids et Mesures. (2007) The international system of units (SI). [Online].
Available: http://www.bipm.org/utils/common/pdf/si brochure 8 en.pdf [12 Nov. 2007].
[4] EPCglobal. (2007, 28 Sep.) EPCglobal Tag Data Standards version 1.3.1. [Online]. Available: http:
//www.epcglobalinc.org/standards technology/EPC TagDataSpecification11Rev126ca.pdf [12
Nov. 2007].
[5] Institute of Electrical and Electronics Engineers. (2007) IEEE technical societies/councils. [Online].
Available: http://www.ieee.org/portal/pages/tab/society.html [12 Nov. 2007].
[6] J. Landt, “The history of RFID,” Potentials, IEEE, vol. 24, no. 4, pp. 8–11, 2005.
[7] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architec-
ture for a ubiquitous item identification network,” in Seventh International Conference on Ubiquitous
computing, Tokyo, Japan, 11-14 Sept. 2005.
[8] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architecture
for a ubiquitous RFID sensing network,” in 2nd International Conference on Intelligent Sensors, Sensor
Networks and Information Processing (ISSNIP), Melbourne, Australia, 5-8 Dec. 2005.
[9] J. Landt, and B. Catlin, “Shrouds of time, the history of RFID,” AIM Publication, 2001.
[10] D. B. Harris, and P. Alto, “Radio transmission systems with modulatable passive responder,” U.S.
Patent Office, Patent 2,927,321, 1960.
[11] R. F. Harrington, “Small resonant scatterers and their use for field measurements,” Microwave
Theory and Techniques, IEEE Transactions on, vol. 10, no. 3, pp. 165–174, 1962.
[12] R. Harrington, “Electromagnetic scattering by antennas,” Antennas and Propagation, IEEE Transac-
tions on [legacy, pre - 1988], vol. 11, no. 5, pp. 595–596, 1963.
[13] P. H. Cole, “Electronic surveillance system,” Australian Patent No. 444,838, U.S. Patent No.
3,706,094, 1972.
Page 333
Bibliography
[14] A. R. Koelle, S. W. Depp, and R. W. Freyman, “Short-range radio-telemetry for electronic identi-
fication, using modulated RF backscatter,” Proceedings of the IEEE, vol. 63, no. 8, pp. 1260–1261,
1975.
[15] K. Finkenzeller, RFID handbook, 2nd ed. John Wiley and Sons, 2003.
[16] International Standards Organization, “ISO/IEC 19762-3: information technology - automatic
identification and data capture (AIDC) techniques - harmonized vocabulary - part 3: radio fre-
quency identification (RFID),” 2005.
[17] U. Karthaus, and M. Fisher, “Fully integrated passive UHF RFID transponder IC with 16.7µ W
minimum RF input power,” Solid-State Circuits, IEEE Journal of, pp. 1602 – 1608, 2003.
[18] R. Glidden, C. Bockorick, S. Cooper, C. Diorio, D. Dressler, V. Gutnik, C. Hagen, D. Hara, T. Hass,
T. Humes, J. Hyde, R. Oliver, O. Onen, A. Pesavento, K. Sundstrom, and M. Thomas, “Design
of ultra-low-cost UHF RFID tags for supply chain applications,” Communications Magazine, IEEE,
vol. 42, no. 8, pp. 140–151, 2004.
[19] S. Lewis. (2004, Jan.) A basic introduction to RFID technology and its use in the supply
chain. [Online]. Available: http://www.printronix.com/library/assets/public/case-studies/
rfid-laran-white-paper-english.pdf [12 Nov. 2007].
[20] G. S. Pope, M. Y. Loukine, D. M. Hall, and P. H. Cole, “Innovative systems design for 13.56 MHz
RFID,” in Wireless and Portable Design Conference, Burlington, Massachusetts, 1997, pp. 240–245.
[21] Philips Semiconductor, TAGSYS, and Texas Instruments Inc. (2004, July 2004) Item-level
visibility in the pharmaceutical supply chain: a comparison of HF and UHF RFID
technology. [Online]. Available: http://www.tagsysrfid.com/modules/tagsys/upload/news/
TAGSYS-TI-Philips-White-Paper.pdf [12 Nov. 2007].
[22] R. Want, “RFID - a key to automating everything,” Scientific American, vol. 290, no. 1, pp. 56–65,
2004.
[23] G. Goth, “RFID: not quite prime time, but dawdle at your own risk,” IEEE Distributed Systems
Online, vol. 6, no. 2, 2005.
[24] C. R. Nave. (2007) Transparency of water in the visible range. [Online]. Available:
http://hyperphysics.phy-astr.gsu.edu/hbase/chemical/watabs.html [12 Nov. 2007].
[25] P. H. Cole, and D. W. Engels, “Auto-ID - 21st century supply chain technology,” in AEEMA Na-
tional Conference, Melbourne Exhibition Centre, Melbourne, Australia, 2002.
[26] R. Angeles, “RFID technologies: supply-chain applications and implementation issues,” Informa-
tion Systems Management, vol. 22, no. 1, pp. 51–65, 2005.
Page 334
Bibliography
[27] T. J. Becker. (2000, 12 Dec.) Putting a price on supply chain problems: study links supply chain
glitches with falling stock prices. [Online]. Available: http://www.gtresearchnews.gatech.edu/
newsrelease/CHAINR.html [12 Nov. 2007].
[28] M. N. Huhns, and L. M. Stephens, “Automating supply chains,” Internet Computing, IEEE, vol. 5,
no. 4, pp. 90–93, 2001.
[29] D. Luckett, “The supply chain,” BT Technology Journal, vol. 22, no. 3, pp. 50–55, 2004.
[30] K. M. Penttila, D. W. Engels, and M. A. Kivikoski, “Radio frequency identification systems in
supply chain management,” International Journal of Robotics & Automation, vol. 19, no. 3, pp. 143–
151, 2004.
[31] R. Weinstein, “RFID: a technical overview and its application to the enterprise,” IT Professional,
vol. 7, no. 3, pp. 27–33, 2005.
[32] Auto-ID Lab. (2007) History of Auto-ID Labs. [Online]. Available: http://autoidlab.eleceng.
adelaide.edu.au/ [12 Nov. 2007].
[33] International Standards Organization, “ISO/IEC 18000-3:2004: information technology - radio fre-
quency identification for item management - Part 3: parameters for air interface communications
at 13.56 MHz,” 2004.
[34] International Standards Organization, “ISO/IEC 15693-1:2000: identification cards - contactless
integrated circuit(s) cards - vicinity cards (Part 1 - 3),” 2000-2001.
[35] EPCglobal. (2007, 17 Oct.) EPC radio-frequency identity protocols class-1 generation-2 UHF
RFID protocol for communications at 860 MHz - 960 MHz version 1.1.0. [Online]. Available:
http://www.epcglobalinc.org/standards/uhfc1g2 [12 Nov. 2007].
[36] D. Barlas. (2003, 4 June) Wal-Mart’s RFID mandate. [Online]. Available: http://www.line56.
com/articles/default.asp?ArticleID=4710 [1 Dec. 2006].
[37] Department of Defence United States of America. (2007, 26 Oct.) Radio frequency identification
(RFID). [Online]. Available: http://www.acq.osd.mil/log/rfid/index.htm [12 Nov. 2007].
[38] S. Sarma, “Integrating RFID,” Queue, vol. 2, no. 7, pp. 50–57, 2004.
[39] U.S. Food and Drug Administration. (2004, 15 Nov.) FDA announces new initiative to protect
the U.S. drug supply through the use of radiofrequency identification technology. [Online].
Available: http://www.fda.gov/bbs/topics/news/2004/NEW01133.html [12 Nov. 2007].
[40] S. Sarma. (2001, 1 Nov.) Towards the 5 cents tag. [Online]. Available: http://www.autoidlabs.
org/uploads/media/mit-autoid-wh-006.pdf [12 Nov. 2007].
[41] EPCglobal. (2007, 10 Sep.) The EPCglobal Architecture Framework version 1.2. [Online].
Available: http://www.epcglobalinc.org/standards/architecture/ [12 Nov. 2007].
Page 335
Bibliography
[42] EPCglobal. (2005, 15 Sep.) The Application Level Events (ALE) Specification version 1.0. [Online].
Available: http://www.epcglobalinc.org/standards/ale [12 Nov. 2007].
[43] EPCglobal. (2007, 21 Sep.) EPC Information Services (EPCIS) version 1.0.1 specification. [Online].
Available: http://www.epcglobalinc.org/standards/epcis [12 Nov. 2007].
[44] EPCglobal. (2005, 4 Oct.) Object Naming Service (ONS) version 1.0. [Online]. Available:
http://www.epcglobalinc.org/standards/ons [12 Nov. 2007].
[45] EPCglobal. (2007, 31 May) Reader Management (RM) Standard version 1.0.1. [Online]. Available:
http://www.epcglobalinc.org/standards/rm [12 Nov. 2007].
[46] EPCglobal. (2006, 21 Jun.) Reader Protocol Standard version 1.1. [Online]. Available:
http://www.epcglobalinc.org/standards/rp [12 Nov. 2007].
[47] Federal Communications Commision, “Title 47, telecommunication, chapter 1, Part 15, radio fre-
quency devices,” 2001.
[48] ETSI, “EN 302 208-1 v1.1.1: electromagnetic compatibility and radio spectrum matters (ERM);
radio frequency identification equipment operating in the band 865 MHz to 868 MHz with power
levels up to 2W,” European Standard (Telecommunications Series), 2004.
[49] International Organization for Standardization. (2007) ISO - International Organization for
Standardization. [Online]. Available: http://www.iso.org/iso/en/ISOOnline.frontpage [12 Nov.
2007].
[50] EPCglobal. (2007) EPCglobal homepage. [Online]. Available: http://www.epcglobalinc.org/
home [12 Nov. 2007].
[51] International Standards Organization, “ISO 11784:1996: Radio-frequency identification of animals
- code structure,” 1996.
[52] International Standards Organization, “ISO 11785:1996: Radio-frequency identification of animals
- technical concept,” 1996.
[53] International Standards Organization, “ISO/IEC 18000-6:2004: information technology - radio fre-
quency identification for item management - Part 6: parameters for air interface communications
at 860 MHz to 960 MHz,” 2004.
[54] EPCglobal. (2003, 1 Feb.) 13.56 MHz ISM band class 1 radio frequency identification
tag interface specification: candidate recommendation, version 1.0.0. [Online]. Avail-
able: http://www.epcglobalinc.org/standards/specs/13.56 MHz ISM Band Class 1 RFID Tag
Interface Specification.pdf [12 Nov. 2007].
[55] B. Jamali, “High-performance RFID System,” Ph.D. dissertation, University of Adelaide, 2005.
Page 336
Bibliography
[56] K. S. Leong, M. L. Ng, and P. H. Cole, “The reader collision problem in RFID systems,” in IEEE
2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless
Communications, Beijing, China, 8-12 Aug. 2005.
[57] K. S. Leong, M. L. Ng, and P. H. Cole, “Positioning analysis of multiple antennas in a dense
RFID reader environment,” in 2006 International Symposium on Applications and the Internet (SAINT)
Workshop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Ari-
zona, USA, 23-27 Jan. 2006.
[58] T. S. Rappaport, Wireless communications - principles and practice, 2nd ed. Prentice Hall, 2002.
[59] ETSI, “EN 302 208-1 v1.2.1: electromagnetic compatibility and radio spectrum matters (ERM);
radio frequency identification equipment operating in the band 865 MHz to 868 MHz with power
levels up to 2W,” European Standard (Telecommunications Series), 2007.
[60] D. Kotz, C. Newport, R. S. Gray, J. Liu, Y. Yuan, and C. Elliott, “Experimental evaluation of wire-
less simulation assumptions,” in Proceedings of the 7th ACM international symposium on Modeling,
analysis and simulation of wireless and mobile systems. Venice, Italy: ACM Press, 2004, pp. 78–82.
[61] D. Eckhardt, and P. Steenkiste, “Measurement and analysis of the error characteristics of an in-
building wireless network,” in Conference proceedings on Applications, technologies, architectures, and
protocols for computer communications. Palo Alto, California, United States: ACM Press, 1996, pp.
243–254.
[62] K. Pawlikowski, H.-D. J. Jeong, and J.-S. R. Lee, “On credibility of simulation studies of telecom-
munication networks,” Communications Magazine, IEEE, vol. 40, no. 1, pp. 132–139, 2002.
[63] K. S. Leong, M. L. Ng, and P. H. Cole, “Operational considerations in simulation and deployment
of RFID systems,” in 17th International Zurich Symposium on Electromagnetic Compatibility, Singa-
pore, 27 Feb. - 3 Mar. 2006.
[64] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of the threshold of second carrier sensing in
RFID deployment,” in 2006 International Symposium on Applications and the Internet (SAINT) Work-
shop, RFID and Extended Network: Deployment of Technologies and Applications, Hiroshima, Japan,
15-19 Jan. 2007.
[65] Workgroup 4 - Radio Frequency Identification for Item Managment (ISO/IEC JTC 1/SC 31/WG4),
“Regulatory status for using RFID in the UHF spectrum,” 2005.
[66] H. Hashemi, “The indoor radio propagation channel,” Proceedings of the IEEE, vol. 81, no. 7, pp.
943–968, 1993.
[67] G. E. Athanasiadou, and A. R. Nix, “A novel 3-D indoor ray-tracing propagation model: the path
generator and evaluation of narrow-band and wide-band predictions,” Vehicular Technology, IEEE
Transactions on, vol. 49, no. 4, pp. 1152–1168, 2000.
Page 337
Bibliography
[68] R. Valenzuela, “A ray tracing approach to predicting indoor wireless transmission,” in Vehicular
Technology Conference, 1993 IEEE 43rd, 1993, pp. 214–218.
[69] S. Y. Seidel, and T. S. Rappaport, “914 MHz path loss prediction models for indoor wireless com-
munications in multifloored buildings,” Antennas and Propagation, IEEE Transactions on, vol. 40,
no. 2, pp. 207–217, 1992.
[70] J.-F. LaFortune, and M. Lecours, “Measurement and modeling of propagation losses in a building
at 900 MHz,” Vehicular Technology, IEEE Transactions on, vol. 39, no. 2, pp. 101–108, 1990.
[71] J. Mitsugi, “Two stage carrier sensing to protect communication links in UHF RFID system,” to be
published, 2006.
[72] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Synchronisation of RFID readers for dense RFID
reader environments,” in 2006 International Symposium on Applications and the Internet (SAINT)
Workshop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Ari-
zona, USA, 23-27 Jan. 2006.
[73] Australian Radiation Protection and Nuclear Safety Agency, “Radiation protection standard -
maximum exposure levels to radiofrequency fields - 3 kHz to 300 GHz,” Radiation Protection Series
No. 3, 2002.
[74] J. Rendon. (2005, 18 Jan.) Wal-Mart touts RFID results. [Online].
Available: http://searchmobilecomputing.techtarget.com/originalContent/0,289142,sid40
gci1048680,00.html?bucket=NEWS [12 Nov. 2007].
[75] D. W. Engels, and S. E. Sarma, “The reader collision problem,” in Systems, Man and Cybernetics,
2002 IEEE International Conference on, vol. 3, 2002, p. 6 pp. vol.3.
[76] B. Carbunar, M. K. Ramanathan, M. Koyuturk, C. Hoffmann, and A. Grama, “Redundant reader
elimination in RFID systems,” in Second Annual IEEE Communications Society Conference on Sensor
and Ad Hoc Communications and Networks (SECON), Santa Clara, California, USA, 2005.
[77] K. S. Leong, M. L. Ng, and P. H. Cole, “Dual-frequency antenna design for RFID application,”
in 21st International Technical Conference on Circuits/Systems, Computers and Communications (ITC-
CSCC 2006), Chiang Mai, Thailand, 10-13 July 2006.
[78] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Dense RFID reader deployment in Europe using
synchronization,” Journal of Communications, vol. 1, no. 7, pp. 9–16, 2006.
[79] K. S. Leong, M. L. Ng, and P. H. Cole, “RFID reader synchronisation,” in RFID Handbook: Applica-
tions, Technology, Security, and Privacy, S. Ahson and M. Ilyas, Eds. CRC, 2008.
[80] T. Skeie, S. Johannessen, and O. Holmeide, “Highly accurate time synchronization over switched
Ethernet,” in Emerging Technologies and Factory Automation, 2001. Proceedings. 2001 8th IEEE Inter-
national Conference on, 2001, pp. 195–204 vol.1.
Page 338
Bibliography
[81] CISC, “Consolidated proposal on synchronization,” ETSI Electromagnetic Compatibility and Radio
Spectrum Matteres, Task Group 34, RF Identification Devices, ETSI ERMTG34#12 012, 2006.
[82] Impinj, “RFID operation in Europe,” ETSI Electromagnetic Compatibility and Radio Spectrum Mat-
teres, Task Group 34, RF Identification Devices, ETSI ERMTG34#12 11r1, 2006.
[83] Intermec, “Smart Listen Before Talk,” ETSI Electromagnetic Compatibility and Radio Spectrum Mat-
teres, Task Group 34, RF Identification Devices, ETSI ERMTG34#12 010, 2006.
[84] RFID Journal. (2006, 10 Oct.) ETSI tests show EPC scaleable in Europe. [Online]. Available:
http://www.rfidjournal.com/article/articleview/2712/1/1/ [12 Nov. 2007].
[85] C. A. Balanis, Antenna theory: analysis and design, 3rd ed. John Wiley and Sons, 2005.
[86] R. C. Johnson, and H. Jasik, Antenna engineering handbook, 3rd ed. McGraw-Hill, 1993.
[87] C. A. Balanis, Antenna theory: analysis and design, 2nd ed. John Wiley and Sons, 1997.
[88] H. A. Wheeler, “The Radiansphere around a Small Antenna,” Proceedings of the IRE, vol. 47, no. 8,
pp. 1325–1331, 1959.
[89] K. Eshragain, “Electromagentic traffic sensing and surveillance,” Ph.D. dissertation, University of
Adelaide, 1980.
[90] I. of Electrical and E. Engineers, “Ieee standards definitions of terms for antennas,” Tech. Rep.,
1993.
[91] Cushcraft Corporation, “Antenna properties and their impact on wireless system performance,”
White Paper, 2002.
[92] J. Bernhard, and E. Michielssen, “Part B: antenna technology and numerical mehtods: antenna
parameters, various generic antenna and feed systems, and available software,” in Handbook of
Antennas in Wireless Communications, L. C. Godara, Ed. CRC Press LLC, 2002.
[93] R. J. Smith, Circuits, devices, and systems, 3rd ed. John Wiley & Sons, 1976.
[94] Texas Instruments. (2006, Oct.) UHF Gen2 STRAP. [Online]. Available: http://www.ti.com/rfid/
docs/manuals/pdfSpecs/RI-UHF-STRAP DataSheet.pdf [12 Nov. 2007].
[95] Texas Instruments, “HF antenna design notes,” Technical Application Report, 2003.
[96] P. V. Nikitin, and K. V. S. Rao, “Performance limitations of passive UHF RFID systems,” in Anten-
nas and Propagation Society International Symposium 2006, IEEE, 2006, pp. 1011–1014.
[97] M. L. Ng, K. S. Leong, and P. H. Cole, “Analysis of constraints in small UHF RFID tag design,”
in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for
Wireless Communications, Beijing , China, 8-12 Aug. 2005.
Page 339
Bibliography
[98] D. C. Ranasinghe, M. L. Ng, K. S. Leong, and P. H. Cole, “Small UHF RFID label antenna design
and limitations,” in IEEE International Workshop on Antenna technology : Small Antennas and Novel
Metamaterials, New York, 6-8 Mar. 2006.
[99] S. R. Best, “A discussion on the quality factor of impedance matched electrically small wire an-
tennas,” Antennas and Propagation, IEEE Transactions on, vol. 53, no. 1, pp. 502–508, 2005.
[100] R. M. Fano, “Theoretical limitations on the broadband matching of arbitrary impedances,” Re-
search Laboratory of Electronics, Massachusetts Institute of Technology, Technical Report No. 41, 1948.
[101] Cushcraft Corporation, “Antenna polarization considerations in wireless communications sys-
tems,” White Paper, 2002.
[102] Texas Instruments. (2005, Nov.) Tag-it HF-I standard transponder IC - reference guide. [Online].
Available: http://www.ti.com/rfid/docs/manuals/refmanuals/HF-IStandardICRefGuide.pdf
[12 Nov. 2007].
[103] NXP, “SL2 ICS20 Smart Label IC Short Form Specification,” 2002.
[104] Ansoft Corporation, “User’s guide - High Frequency Structure Simulator,” 2003.
[105] T. H. Hubing, “Survey of numerical electromagnetic modeling techniques,” Technical Report for
Intel Corporation, TR91-1-001.3, 1991.
[106] Ansoft Corporation, “Introduction to scripting in HFSS,” 2003.
[107] V. C. Ramasami. (2007) HFSS MATLAB scripting API. [Online]. Available: http://www.ittc.ku.
edu/∼rvc/projects/hfssapi/doc/hfss-matlab-api.html [18 Jan. 2007].
[108] T-Tech Inc. (2007) PCB prototyping, autoSketch, quick circuit system, ISOPro software. [Online].
Available: http://www.t-tech.com/default.asp [12 Nov. 2007].
[109] Y. Lee, “Antenna circuit design for RFID applications,” Microchip Technology Inc. Application Note,
vol. AN710, 2003.
[110] Texas Instruments, “HF antenna cookbook,” Technical Application Report, 2004.
[111] S. S. Basat, K. Lim, J. Laskar, and M. M. Tentzeris, “Design and modeling of embedded 13.56 MHz
RFID antennas,” in Antennas and Propagation Society International Symposium, 2005 IEEE, vol. 4B,
2005, pp. 64–67 vol. 4B.
[112] S. C. Q. Chen, and V. Thomas, “Optimization of inductive RFID technology,” in Electronics and the
Environment, 2001. Proceedings of the 2001 IEEE International Symposium on, 2001, pp. 82–87.
[113] H. Greenhouse, “Design of planar rectangular microelectronic inductors,” Parts, Hybrids, and Pack-
aging, IEEE Transactions on, vol. 10, no. 2, pp. 101–109, 1974.
[114] Zork. (2007) The alternative to cork. [Online]. Available: http://www.zork.com.au/ [12 Nov.
2007].
Page 340
Bibliography
[115] ATMEL Corporation, “Datasheet - ATA5558 - 1kbit R/W IDIC with deterministic anticollision,”
2005.
[116] B. Mennecke, and A. Townsend. (2005, May) Radio frequency identification tagging as a
mechanism of creating a viable producer’s brand in the cattle industry. [Online]. Available:
http://www.card.iastate.edu/publications/DBS/PDFFiles/05mrp8.pdf [12 Nov. 2007].
[117] Allflex Australia. (2007) Allflex 36T NLIS tags. [Online]. Available: http://www.allflex.com.au/
nlis tags.htm [12 Nov. 2007].
[118] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation on the deployment of HF and UHF RFID
tag in livestock identification,” in IEEE Antennas and Propagation Society International Symposium,
Honolulu, Hawaii, USA, 10-15 Jun. 2007.
[119] K. S. Leong, M. L. Ng, and P. H. Cole, “HF and UHF RFID tag design for pig tagging,” in 11th
Biennial Conference of the Australasian Pig Science Association (APSA), Brisbane, Australia, 25-28 Nov.
2007.
[120] K. V. S. Rao, P. V. Nikitin, and S. F. Lam, “Antenna design for UHF RFID tags: a review and a
practical application,” Antennas and Propagation, IEEE Transactions on, vol. 53, no. 12, pp. 3870–
3876, 2005.
[121] M. L. Ng, K. S. Leong, D. M. Hall, and P. H. Cole, “A small passive UHF RFID tag for livestock
identification,” in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC
Technologies for Wireless Communications, Beijing, China, 8-12 Aug. 2005.
[122] Alien Technologies. (2007) 915MHz EPC Class 1 RFID tags. [Online]. Available: http:
//www.alientechnology.com/products/documents/alien 915mhz 128 bit.pdf [15 Jan. 2007].
[123] M. L. Ng, K. S. Leong, and P. H. Cole, “A small passive UHF RFID tag for metallic item identifica-
tion,” in 21st International Technical Conference on Circuits/Systems, Computers and Communications
(ITC-CSCC 2006), Chiang Mai, Thailand, 10-13 July 2006.
[124] K. S. Leong, M. L. Ng, and P. H. Cole, “Miniaturization of dual-frequency RFID antenna with
high frequency ratio,” in IEEE Antennas and Propagation Society International Symposium, Honolulu,
Hawaii, USA, 10-15 Jun. 2007.
[125] K. L. Wong, Compact and broadband microstrip antennas. John Wiley and Sons, 2002.
[126] W. Menzel, M. Al-Tikriti, and M. B. E. Lopez, “A common aperture, dual-frequency printed an-
tenna (900 MHz and 60 GHz),” Electronics Letters, vol. 37, no. 17, pp. 1059 – 1060, 2001.
[127] J. Collins. (2004, 1 Sep.) New two-frequency RFID system, RFID Journal. [Online]. Available:
http://www.rfidjournal.com/article/articleview/1105/1/1/ [12 Nov. 2007].
Page 341
[128] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of RF cable effect on RFID tag antenna
impedance measurement,” in IEEE Antennas and Propagation Society International Symposium, Hon-
olulu, Hawaii, USA, 10-15 Jun. 2007.
[129] O. Staub, J.-F. Zrcher, and A. Skrivervik, “Some considerations on the correct measurement of
the gain and bandwidth of electrically small antennas,” Microwave and Optical Technology Letters,
vol. 17, no. 3, pp. 156–160, 1998.
[130] J. DeMarinis, “The antenna cable as a source of error in EMI measurements,” in Electromagnetic
Compatibility, 1988. Symposium Record. IEEE 1988 International Symposium on, 1988, pp. 9–14.
[131] C. Icheln, M. Popov, P. Vainikainen, and S. He, “Optimal reduction of the influence of RF feed
cables in small antenna measurements,” Microwave and Optical Technology Letters, vol. 25, no. 3,
pp. 194–196, 2000.
[132] C. Icheln, J. Krogerus, and P. Vainikainen, “Use of balun chokes in small-antenna radiation mea-
surements,” Instrumentation and Measurement, IEEE Transactions on, vol. 53, no. 2, pp. 498–506,
2004.
[133] Z. N. Chen, N. Yang, Y.-X. Guo, and M. Y. W. Chia, “An investigation into measurement of handset
antennas,” Instrumentation and Measurement, IEEE Transactions on, vol. 54, no. 3, pp. 1100–1110,
2005.
[134] S. Kahng, “A stack of metal rings for reducing common-mode current on a wire passing through
an aperture,” IEICE Trans Commun, vol. E88-B, no. 9, pp. 3819–3822, 2005.
[135] S. Saario, D. V. Thiel, J. W. Lu, and S. G. O’Keefe, “An assessment of cable radiation effects on
mobile communications antenna measurements,” in Antennas and Propagation Society International
Symposium, 1997. IEEE., 1997 Digest, vol. 1, 1997, pp. 550–553 vol.1.
[136] R. N. Simons, Coplanar waveguide circuits, components, and systems. John Wiley & Sons, 2001.
Page 342