Studies on Fade Mitigation Control for Microwave …gssst.iitkgp.ac.in/GSProjects/SFM/reports/SFM_main...Date:November 10, 2010 Acknowledgement I would like to thank my project supervisor

'

&

$

%

Project Report

on

Studies on Fade Mitigation Control for

Microwave Satellite Signal Propagation

by

Jayeeta Saha

(Roll No. 10GS7001)

under the guidance of

Dr.Suvra Sekhar Das

G S Sanyal School Of Telecommunications

Indian Institute of Technology Kharagpur

Kharagpur - 721302, India

Date:November 10, 2010

Acknowledgement

I would like to thank my project supervisor Prof. Suvra Sekhar Das for inspiring and

motivating me to develop new ideas and implementing them.I am grateful to Prof.

Kalyan Kumar Bandyopadhyay for helpful suggestions during course work.Last but

not the least,I would like to thankmy team-mates SantanuMondal(ECE,B.Tech,06EC1013)

and G Srinivas Sagar(ECE,M.Tech,08EC6414),both graduated from this institute in this

year for their important contribution to this project.

Abstract

There is worldwide interest, including ISRO, to use higher than C band spectrum

in future Satellite Communication Systems. They offer several advantages for Satel-

lite Communications over C band like, spectrum availability, reduced terrestrial Inter-

ference potential and reduced equipment size. However, higher spectrum bands are

more susceptible to tropospheric impairment that can severely degrade service qual-

ity. If estimation of such impairments can be done then proper Mitigation Techniques

can be implemented to improve service quality. Typical Communication Systems use a

margin to overcome the channel fades. Channel fades can occur from several sources.

Use of such static margin is not advantageous. However if the channel fades can

be predicted then transmission signal may be designed so as to avoid the fades /

take advantage of good channel conditions. Such types of systems are known as link

adaptation systems where link level parameters are dynamically adjusted in order to

maximize the data rates over a certain period of time. Several results exists for ter-

restrial cellular communication systems, but these have not much been experimented

for Satellite Communication systems and especially above C band. The objective of

project is to counteract the propagation effects at the physical layer level. Some of

the techniques are power control, adaptive waveform, diversity and layer 2. These

techniques allow systems with small static margin to be designed, while overcoming

the propagation impairments. Among those techniques, adaptive modulation/coding

are of high interest as they allow the performance of individual links to be optimized

and the transmission characteristics to be adapted to the propagation channel condi-

tions and to the service requirements for the given link. We also incorporate the delay

compensation strategies into the system.

i

Contents

Abstract i

List of Figures v

List of Tables viii

1 Introduction 1

1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 State of Art in Satellite Communication . . . . . . . . . . . . . . . . . . . 3

1.4 Problem area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.4.1 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Frame Work 6

3 System Description 7

3.1 System Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.2 Channel Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.2.1 Propagation effects and their impact on satellite-earth links . . . 8

3.2.2 Link Budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3.3 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3.4 Decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3.5 Fade Mitigation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3.5.1 Power control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3.5.2 Adaptive waveform . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.5.3 Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.5.4 Layer 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.6 Link adaptation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.6.1 BER performance for different modulation schemes . . . . . . . . 19

4 Fade Mitigation Techniques 21

4.1 Propagation effects and their impact on satellite-earth links . . . . . . . . 21

ii

CONTENTS

4.1.1 Link Budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

4.2 Fade Mitigation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4.2.1 Power control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

4.2.2 Adaptive waveform . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4.2.3 Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4.2.4 Layer 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.3 Link adaptation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30


5 Implementation of FMT 33

5.1 FMT control logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

5.2 Implementation of FMT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

5.3 Description of simulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

5.3.1 CRC Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

5.3.2 Error Control Coding . . . . . . . . . . . . . . . . . . . . . . . . . . 35

5.3.3 Modulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

5.3.4 Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

5.3.5 Automatic Gain Control (AGC) . . . . . . . . . . . . . . . . . . . . 36

5.3.6 demodulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5.3.7 Decoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5.3.8 CRC Decoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

6 Channel 38

6.1 Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

6.2 Channel considered in simulation . . . . . . . . . . . . . . . . . . . . . . . 39

7 Detection 41

7.1 Methods from the literature . . . . . . . . . . . . . . . . . . . . . . . . . . 42

7.2 Fade detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

7.2.1 Fade detection using CRC . . . . . . . . . . . . . . . . . . . . . . . 42

7.2.2 Detection using Embedded pilot . . . . . . . . . . . . . . . . . . . 44

7.2.3 Continuous pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

7.2.4 Detection using Embedded pilot . . . . . . . . . . . . . . . . . . . 44

7.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45


7.3.2 Performance of the system for the collected data . . . . . . . . . . 46

7.3.3 PER performance for different SNR values . . . . . . . . . . . . . 49

7.3.4 Performance of the FMT system with time . . . . . . . . . . . . . 50

7.3.5 SNR estimation accuracy with No Back Off . . . . . . . . . . . . . 50

7.3.6 SNR estimation accuracy with Symmetric Back Off . . . . . . . . 50

iii

CONTENTS

7.3.7 SNR estimation accuracy with Asymmetric Back Off . . . . . . . 51

7.3.8 SNR estimation accuracy with Adaptive Back Off . . . . . . . . . 52

8 Decision 55

8.1 Decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

8.1.1 Detection margin and Hysteresis . . . . . . . . . . . . . . . . . . . 56

8.2 Decision making algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 57

8.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59


8.3.2 Performance of the system for the collected data . . . . . . . . . . 59

8.3.3 PER performance for different SNR values . . . . . . . . . . . . . 63

8.3.4 Performance of the FMT system with time . . . . . . . . . . . . . 63

9 Delay compensation 65

9.1 Delay calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

9.2 Delay Compensation strategies . . . . . . . . . . . . . . . . . . . . . . . . 67

9.3 Delay Compensation flow chart . . . . . . . . . . . . . . . . . . . . . . . . 68

9.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

9.4.1 SNR estimation with CRC and delay compensation . . . . . . . . 69

9.4.2 SNR estimation with Continuous pilot and delay compensation . 72

9.4.3 SNR estimation with Distributed pilot and delay compensation . 74

10 Results 76

10.1 Without SNR moving average . . . . . . . . . . . . . . . . . . . . . . . . . 76

10.1.1 Without SNR moving average and adaptive back off . . . . . . . 76

10.1.2 Without SNR moving average and no adaptive back off . . . . . . 77

10.2 With SNR moving average . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

10.2.1 With SNR moving average and adaptive back off . . . . . . . . . 78

10.2.2 With SNR moving average and no adaptive back off . . . . . . . 79

11 Updated Results 80

11.1 Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

12 Practical Implementation with Modem (SRM6100) 85

12.1 Introduction of Modem-SRM6100 . . . . . . . . . . . . . . . . . . . . . . . 85

12.2 Specifications of Modem-SRM6100 . . . . . . . . . . . . . . . . . . . . . . 85

12.3 Experiments done with the modem SRM6100 . . . . . . . . . . . . . . . . 86

12.3.1 Loop Back Bench Test . . . . . . . . . . . . . . . . . . . . . . . . . 86

12.3.2 Configuration setting . . . . . . . . . . . . . . . . . . . . . . . . . . 87

12.4 Limitations of modem SRM6100 . . . . . . . . . . . . . . . . . . . . . . . 87

iv

CONTENTS

13 Plan of the experiment had to be done at SAC,Ahmedabad 89

13.1 Objective of the experiment . . . . . . . . . . . . . . . . . . . . . . . . . . 89

13.2 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

13.3 Experimental Set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

13.4 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

13.5 Resources required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

13.6 Expected Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

14 Conclusion 93

14.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

14.2 Future scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

A List of Abbreviations 95

Bibliography 99

v

List of Figures

2.1 the basic frame work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3.1 FMT System Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.2 Various ways of power control . . . . . . . . . . . . . . . . . . . . . . . . 15

3.3 site diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.4 EbNo verses probability of error curves . . . . . . . . . . . . . . . . . . . 19

3.5 SNR verses BER curves taken from [7] . . . . . . . . . . . . . . . . . . . . 20

4.1 Various ways of power control . . . . . . . . . . . . . . . . . . . . . . . . 26

4.2 site diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.3 EbNo verses probability of error curves . . . . . . . . . . . . . . . . . . . 31

4.4 SNR verses BER curves taken from [7] . . . . . . . . . . . . . . . . . . . . 32

5.1 System block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

5.2 Block diagram of FCM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

5.3 variation of channel with time . . . . . . . . . . . . . . . . . . . . . . . . . 36

7.1 SNR detection using CRC method . . . . . . . . . . . . . . . . . . . . . . 43

7.2 SNR detection using Continuous Pilot method . . . . . . . . . . . . . . . 45

7.3 SNR verses probability of error curves . . . . . . . . . . . . . . . . . . . . 46

7.4 Change of modulation and coding with time (date 6th) . . . . . . . . . . 47

7.5 Change of modulation and coding with time (date 5th) . . . . . . . . . . 47

7.6 Change of BER with time using collected data (date 5th) . . . . . . . . . 48

7.7 Change of BER with time using collected data (date 6th) . . . . . . . . . 48

7.8 PER versus SNR curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

7.9 Time versus SNR and change of M and C . . . . . . . . . . . . . . . . . . 50

7.10 SNR estimation accuracy between Distributed Pilot and Continuous Pi-

lot methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

7.11 SNR estimation accuracy between Continuous Pilot and CRC methods . 51

7.12 SNR estimation accuracy between Distributed Pilot and CRCmethods . 52

7.13 SNR estimation accuracy between Distributed Pilot, Continuous Pilot

and CRC methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

vi

LIST OF FIGURES

7.14 SNR estimation accuracy between Distributed Pilot, Continuous Pilot

and CRC methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

7.15 SNR estimation accuracy between Distributed Pilot and Continuous Pi-

lot methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

7.16 SNR estimation accuracy between Continuous Pilot and CRC methods . 54

7.17 SNR estimation accuracy between Distributed Pilot and CRCmethods . 54

8.1 FMT control logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

8.2 Decision making flow chart . . . . . . . . . . . . . . . . . . . . . . . . . . 57

8.3 SNR verses probability of error curves . . . . . . . . . . . . . . . . . . . . 59

8.4 change of modulation and coding with time(date 6th) . . . . . . . . . . . 60

8.5 change of modulation and coding with time(date 5th) . . . . . . . . . . . 60

8.6 change of ber with time (with collected data(date 5th)) . . . . . . . . . . 61

8.7 change of ber with time (with collected data(date 6th)) . . . . . . . . . . 62

8.8 PER versus SNR curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

8.9 Time versus SNR and change of M and C . . . . . . . . . . . . . . . . . . 64

9.1 Delay calculation block diagram . . . . . . . . . . . . . . . . . . . . . . . 66

9.2 Delay Compensation flow chart . . . . . . . . . . . . . . . . . . . . . . . . 68

9.3 Delay Compensation flow chart using adaptive back off . . . . . . . . . . 69

9.4 Comparison of SNR and data rate curves with and without back off for

CRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70


CRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

9.6 Comparison of SNR curves with and without back off for CRC . . . . . . 71


continuous pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72


continuous pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

9.9 Comparison of SNR curves with and without back off for Continuous

pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73


distributed pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74


distributed pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

9.12 Comparison of SNR curves with and without back off for Distributed

pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

10.1 Comparison of BLER and SNR curves Without SNR moving average

and adaptive back off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

vii

LIST OF FIGURES

10.2 Comparison of BLER and SNR curves Without SNR moving average

and no adaptive back off . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

10.3 Comparison of Throughput and SNR curves Without SNR moving av-

erage and no adaptive back off . . . . . . . . . . . . . . . . . . . . . . . . 77

10.4 Comparison of BLER and SNR curves With SNR moving average and

adaptive back off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

10.5 Comparison of Throughput and SNR curvesWith SNRmoving average

and adaptive back off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

10.6 Comparison of BLER and SNR curves With SNR moving average and

no adaptive back off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

10.7 Comparison of Throughput and SNR curvesWith SNRmoving average

and no adaptive back off . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

11.1 Throughput and SNR curves of different Back-Off schemes . . . . . . . . 80

11.2 Error performance of different Back-Off schemese . . . . . . . . . . . . . 81

11.3 Throughput performance of different Back-Off schemes . . . . . . . . . . 82

11.4 CDF of BLER for different Back-Off schemes . . . . . . . . . . . . . . . . 82

11.5 CDF of Throughput perforomance of different Back-Off schemes . . . . 83

11.6 Cross-Correlation between the calculated and estimated SNR . . . . . . 84

13.1 plan of the experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

viii

List of Tables

1.1 Radio spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

3.1 Link Analysis for Regenerative Payload(Fair weather) taken from [19] . 11

3.2 Link Analysis for Bent pipe Payload(Fair weather) taken from[19] . . . . 12

4.1 Link Analysis for Regenerative Payload(Fair weather) . . . . . . . . . . . 24

4.2 Link Analysis for Bent pipe Payload(Fair weather) . . . . . . . . . . . . . 25

ix

Chapter 1

Introduction

1.1 Background

In Telecommunication the use of satellite is to provide communication between vari-

ous points on earth. Basically the satellites acts as relays wherein the signals like voice,

video and data are relayed between various earth stations. The basic mechanism of a

communication satellite involves transmitting signals from an earth station to a satel-

lite, the satellite will receive the signal, amplify the signal,and retransmits the signal to

the region of earth where the destination is located. Receiving stations in that particu-

lar region will pick up the signals this completes the whole communication. Satellite

systems operate in microwave and millimeter wave frequency bands,using frequen-

cies between 1 and 50 GHz.

All satellites require radio spectrum. Different parts of the radio spectrum are used

for different radio transmission technologies and applications. Successive world ra-

dio conferences have allocated new frequency bands for commercial satellite services

that now include L,S,C,Ku,Ka,V ,and Q bands. Mobile satellite systems use VHF,

UHF,L,and S bands with carrier frequencies from 137 to 2500 MHz, and GEO satel-

lites use frequency bands extending from 3.2 to 50 GHz. Despite the growth of fiber

optic links with very high capacity, the demand for satellite system continues to in-

crease. The microwave spectrum is usually defined as electromagnetic energy ranging

from approximately 1 GHz to 100 GHz in frequency, but older usage includes lower

frequencies. Most common applications are within the 1 to 40 GHz range.

Satellite communication is started with C band. C band constitutes 6 GHz for uplink

and 4 GHz for downlink. All those satellites that are operating in C band have to

be placed at 2 degrees apart so the Geo Stationary Orbit(GEO) is filled up with the

satellites operating at C band. Therefore the satellites were built for next available

frequency bands, like Ku and Ka bands. There is a continuing demand for ever more

spectrum to allow satellite to provide new services, with high speed access to the inter-

1

Introduction

Band name Frequency RangeL Band 1 to 2 GHzS Band 2 to 4 GHzC Band 4 to 8 GHzX Band 8 to 12 GHzKu Band 12 to 18 GHzK Band 18 to 26.5 GHzKa Band 26.5 to 40 GHzQ Band 30 to 50 GHzU Band 40 to 60 GHzV Band 50 to 75 GHzD Band 110 to 170 GHz

Table 1.1: Radio spectrum

net forcing a move to the Ka band and even higher frequencies. But as the frequency of

operation increases the impairments offered by these frequencies will also increases.

In order to avoid these impairments we need to implement the fade mitigation tech-

niques.

1.2 Motivation

There is worldwide interest, including ISRO, to use higher than C band spectrum in

future Satellite Communication Systems. They offer several advantages for Satellite

Communications over C band like, spectrum availability, reduced terrestrial Inter-

ference potential and reduced equipment size. However, higher spectrum bands are

more susceptible to tropospheric impairment that can severely degrade service qual-

ity. If estimation of such impairments can be done then proper Mitigation Techniques

can be implemented to improve service quality.

Typical Communication Systems use a margin to overcome the channel fades. Chan-

nel fades can occur from several sources. Depending upon the type of fade the margin

may vary. In case of short term fading, more than 40 dB of margin may be required in

Ka band in order to guarantee satisfactory service availability. Use of static margins

takes away a huge amount from the link budget leaving the options of low data rates.

However if the channel fades can be tracked / predicted then transmission signal may

be designed so as to avoid the fades / take advantage of good channel conditions.

Such types of systems are known as link adaptation systems where link level parame-

ters are dynamically adjusted in order to maximize the data rates over a certain period

of time. Several results exists for terrestrial cellular communication systems, but these

have notmuch been experimented for Satellite Communication systems and especially

above C band. So it will be very interesting to implement the link adaptation at higher

2

Introduction

frequency bands in satellite communication.

1.3 State of Art in Satellite Communication

Satellite communication is started with C band, as the GEO orbit is filled with satel-

lites operating in C band, the satellites were built for next available frequency bands

like Ku and Ka. These higher frequency bands offer several advantages like, spec-

trum availability, reduced terrestrial Interference potential and reduced equipment

size. However, higher spectrum bands are more susceptible to tropospheric impair-

ment that can severely degrade service quality.

GSAT-4 is an experimental communication satellite planned for launch in April 2010

by the Indian Space Research Organization using a GSLV rocket. Weighing around

two tonnes, GSAT-4will carry amulti-beamKa-band bent pipe and regenerative transpon-

der and navigation payload in C, L1 and L5 bands.

1.4 Problem area

The constant growth of communication services, both in number of users and amount

of data rate, and the limited available frequency resources at Ku-Band, pushes the

satellite industry to consider implementation of future satellite systems operating at

Ka-band and above where large bandwidths are available. As the frequency of opera-

tion is increased, the attenuation effects of atmospheric gas, clouds and rain and scin-

tillation become more important. These impairments can be compensated by adding a

large amount of power margin but it is not cost efficient to design a large power mar-

gin. So link signal fading must be compensated by other means in order to increase

system availability. Hence use fade mitigation techniques to overcome the fade due

to atmospheric impairments at higher frequency bands. So fade mitigation techniques

have to be incorporated into the system then the system can adapt its physical layer

to the propagation channel variations, optimizing system capacity in clear sky and

reaching the required availability during unfavorable propagation conditions. The

performance of FMT system is dependent not only on the type of the fade mitigation

technique used but also on the type of algorithm which is used to implement the cor-

responding FMT.

Whatever may be the fade mitigation technique used, a control loop is necessary. This

control loop should perform the following functions. first it has to detect the amount

of fade present in the channel, then it has to predict the channel fade a short time

ahead, then it has to take the decision whether to activate FMT or not. If it has to acti-

vate FMT then what kind of FMT it has to select. These are the main things that have

3

Introduction

to be done regularly and based on this information the system should adapt the sys-

tem parameters. These FMT’s should track the signal variations, especially the slow

component (attenuation), and possibly the envelope of fast fluctuations.

Various methods exist to counteract the fade or propagation impairments at the phys-

ical layer level. Some of them are listed below

1. Power control: The power of transmitter is adapted according to the channel

impairments

2. Adaptive Waveform: The modulation scheme and coding is adapted according

to channel impairments.

3. Diversity: Fade is compensated by using a link which has less channel impair-

ments

4. Layer 2: coping with the temporal dynamics of the fade

Fade Mitigation Technique have to be considered and have to be introduced into the

system through the design of a control loop, which aims at mitigating a propagation

event in real time, by adapting some systems parameters : transmitted power, coding,

modulation. The dynamics of the channel is therefore a key element to be taken into

account directly into the definition of FMT control loop. At the Ka-band,propagation

impairments strongly limit the quality and availability of satellite communications.

Adaptive impairment mitigation techniques have to be used in order to improve link

performance. Amongst all other fade mitigation techniques, our interest is to im-

plement adaptive modulation and coding technique. Here we need to estimate the

amount of channel impairments accordingly we have to change themodulation scheme

and coding rate that is if we have less impairments in the channel then we can go for

higher order modulation scheme and coding rate which provides us the greater data

rate satisfying the required BER constraint. Whereas if we have more impairments

then we can go for lower order modulation scheme and coding which gives us less

data rate but we can have the link without failure. So in this way we have to adapt the

modulation scheme and coding according to channel impairments.

1.4.1 Objective

The main objective of the project is to implement the fade mitigation technique in

satellite earth links at Ka band. Amongst all other fade mitigation techniques, our

interest is to implement adaptive modulation and coding technique. Here we need

to estimate the amount of channel impairments accordingly we have to change the

modulation scheme and coding rate. The performance of FMT system with adaptive

coding and modulation has to be studied. Due to the inherent delay of the satellite

4

Introduction

system, the channel estimation information is delayed by approximately 240 ms (for a

GEO satellite case). So we need to incorporate Delay compensation strategies into the

system.

5

Chapter 2

Frame Work

The pictorial description of the project frame work is shown in figure 2.1. Description

Figure 2.1: the basic frame work

6

Chapter 3

System Description

3.1 System Description

Satellite systems require radio spectrum. As the GEO orbit is filled up with satellites

operting in C and Ku bands the satellites were built for next higher frequency bands to

offer broader transmission channels for multimedia applications. An increasing num-

ber of new services are being promoted for Ka-band (20/30 GHz) satellite systems,

involving very small aperture terminals (VSAT). At the Ka-band,propagation impair-

ments strongly limit the quality and availability of satellite communications.

In order to avoid the fade we need to go for fade mitigation techniques. FMT sys-

Figure 3.1: FMT System Description

7

System Description

tem description is presented in fig.. 3.1. The FMT system contains three parts one is

channel modeling, second is detection of channel impairments and third is taking the

decision for activating the FMT. Further improvement in system performance can be

done using delay compensation techniques. The channel modeling is to be done by

Jayeeta, the detection of channel impairments using CRC and Embedded pilot is done

by Santanu and the decision and delay compensation are done by me. Whatever the

FMT is used the above control logic should be followed. The literature for the system

is presented below.

3.2 Channel Modeling

First we need to model the channel, to model the channel first of all we need to know

the different kinds of impairments that are there in the atmosphere. Then we need to

find the amount of attenuation presented by each impairment so that we can find the

overall attenuation presented by atmosphere. There are many phenomena that lead to

signal loss on transmission through the earths atmosphere like Atmospheric absorp-

tion, Cloud attenuation, Ionospheric scintillation, Tropospheric scintillation, and Rain

attenuation. The detailed study of these attenuation models has been presented in ITU

documents[6].

3.2.1 Propagation effects and their impact on satellite-earth links

All radio wave signals have to be transmitted through the atmosphere. These signals

will be effected by the atmospheric impairments. The effects of atmosphere have to be

considered in system design at frequencies above 20 GHz. There are different kinds

of atmospheric impairments like atmospheric absorption, cloud attenuation, Tropo-

spheric scintillation, Low angle fading, Ionospheric scintillation and Rain attenuation.

A brief summary of atmospheric impairments will be presented in this section. The

literature is taken from [2] and [6].

Atmospheric absorption

At microwave frequencies and above, electromagnetic waves interact with molecules

in the atmosphere to cause signal attenuation. At certain frequencies, resonant absorp-

tion occurs and severe attenuation can result.The amount of attenuation is less than 1

dB on most paths below 100 GHz.

8

System Description

Cloud attenuation

Clouds have become an important factor for someKa-band paths and all V-band(50/40

GHz) systems. The difficulty with modeling cloud attenuation is that clouds are of

many types and can exist at many levels.The water droplet concentrations in each

cloud will also vary,and clouds made up of ice crystals cause little attenuation. The

amount of attenuation is 1 and 2 dB at frequencies around 30 GHz

Tropospheric scintillation

Energy from the sun warms the surface of the earth and the resultant convective activ-

ity agitates the boundary layer. This agitation results in turbulent mixing of different

parts of boundary layer, causing small scale variations in refractive index. The rapid

variations in refractive index along the path will lead to fluctuations in the received

signal level these fluctuations are known as tropospheric scintillations.

Low angle fading

When the elevation angle falls below 10 degrees, a second propagation effect becomes

noticeable that is low angle fading. Low angle fading is the same phenomena as mul-

tipath fading in terrestrial paths. A signal transmitted from a satellite arrives at the

earth station receiving antenna via different paths with different phase shifts. On the

combination, the resultant waveformmay be enhanced or attenuated from the normal

clear sky level.

Ionospheric scintillation

Energy from sun causes the ionosphere to grow during the day, increasing the total

electron content (TEC) by two orders of magnitude, or more. The rapid change in

TEC from the daytime to nighttime, which occurs at local sunset in the ionosphere,

that gives rise to irregularities in the ionosphere. These rapid fluctuations are called

ionospheric scintillations.

Rain attenuation

At frequencies above 10 GHz , rain is the dominant propagation phenomenon on satel-

lite links. Rain drops absorb and scatter the electromagnetic waves. In Ku and Ka

bands rain attenuation is almost entirely caused by absorption.At Ka band there is a

small contribution from scattering by large rain drops.Rain is the primary cause of

depolarization. Atmospheric gases and tropospheric scintillation do not cause signal

depolarization.Ionosphere causes the depolarization.Some of the energy in one polar-

ization can cross over to other polarization due to asymmetric particles in the existing

9

System Description

path, which leads to depolarization. Measure of depolarization that is most useful

in analyzing communication system is the cross pole isolation(XPI).It is the ratio of

wanted power to the unwanted power.

10

System Description

3.2.2 Link Budget

Once we come to know the amount of attenuation in the atmosphere we will calculate

the link budget. Link budget is nothing but the tabular method of calculating the

received power and noise power.

Link Analysis for Regenerative Payload(Fair weather)

Link budget is nothing but the tabular method of calculating the received power and

noise power. When we calculate our link budget then we come to know how much

amount of margin we have and accordingly we will go for implementing the fade

mitigation techniques. Link Analysis for Regenerative Payload(Fair weather) has been

presented below.

Parameters Unit NB Terminal WB Terminal

Transmit Frequency GHz 29.6 to 30.2 29.6 to 30.2Receive Frequency GHz 20.6 to 21.6 20.6 to 21.6Antenna size Meter 0.28 offset feed parabolic 0.75 offset feed parabolicUP LINK

SSPA Power output W 5.0 10.0Terminal Antenna gain dB 36.3 45.2Terminal EIRP dBW 42.8 54.35Satellite G/T dB/K 6.0 6.0C/No Available dBHz 62.9 74.4Data rate Kbps 64 2048C/No req for BER of 10−7 dBHz 57.1 72.1Available Link Margin dB 5.8 2.35DOWN LINK

SSPA EIRP dBW 40.0 42.0Terminal Antenna gain dB 33.2 41.76System noise temp K 315 315Terminal G/T dB/K 8.7 17.2C/No Available dBHz 66.8 77.3Data rate Kbps 576 2048C/No req for BER of 10−6 dBHz 63.6 69.1Available Link Margin dB 3.2 8.26

Table 3.1: Link Analysis for Regenerative Payload(Fair weather) taken from [19]

11

System Description

Link Analysis for Bent pipe Payload(Fair weather)

Link Analysis for Bent pipe Payload for Fair weather has been shown in the table.

Parameters Unit NB TerminalTransmit Frequency GHz 29.6 to 30.2Receive Frequency GHz 20.6 to 21.6Antenna size Meter 2.4 CassegrainUP LINKSSPA Power output W 13.0Terminal Antenna gain dB 55.21Terminal EIRP dBW 67.22Satellite G/T dB/K 6.0C/No Available dBHz 88.13DOWN LINKSSPA EIRP dBW 44.31Terminal Antenna gain dB 52.06Terminal G/T dB/K 27.56C/No Available dBHz 89.93Total C/No dBHz 86.33Data rate Mbps 40.00C/No req for BER of 10−6 dBHz 82.02Available Link Margin dB 4.31

Table 3.2: Link Analysis for Bent pipe Payload(Fair weather) taken from[19]

3.3 Detection

The amount of attenuation present in the channel is measured through the measure-

ments. Measurements is nothing but the detection of attenuation. The objective of

the detection function is to quantify the magnitude of a fade event occurring on the

considered link. Three kinds of detection concepts are

1. Open loop detection

2. Closed loop detection

3. Hybrid loop detection

Open-loop Detection

The open-loop detection concept relies on the estimation of uplink (or downlink) im-

pairment from a measurement of the propagation conditions. This measurement can

be carried out in several ways: rain intensity and other meteorological measurements,

12

System Description

sky brightness temperatures measured with a radiometer, radar networks, satellite

imagery or satellite beacon operating at uplink or downlink frequency.

Closed-loop Detection

In the closed-loop detection concept, estimation of the impairment is performed from

the measurement of the overall link performance. Bit Error Rate or Carrier plus Noise

estimations can be carried out by the earth station [6] or by the satellite (if On-Board-

Processing enables so). In the case of a transparent satellite link, a measurement of the

overall link will give information on the total degradation of the propagation channel.

However, it will not identify if the impairment is occurring on the uplink or on the

downlink.

Hybrid-loop Detection

To separate uplink and downlink fade contributions the hybrid-loop detection concept

uses two different measurements, one of them from a beacon and the other from the

link .

3.4 Decision

The objective of decision function is to take a decision whether to activate FMT or not.

If FMT is to be activated what kind of FMT is to be considered. The details of the

decision[1] are presented in later chapter.

3.5 Fade Mitigation Techniques

As the operating frequency is increased, the atmospheric attenuations becomes more

severe. so implementing static margin as the only mean to compensate the propaga-

tion impairments at high frequency bands is not a good task, and it will push towards

the implementation of Fade Mitigation Techniques(FMT).Those techniques allow sys-

tems with rather small static margin to be designed, while overcoming in real time

cloud attenuation, some fraction of rain attenuation,scintillation, and depolarization

events. The review of fade mitigation techniques has been taken from [1].

Making use of FadeMitigation Techniques involves adapting in real time the link bud-

get to the propagation conditions through some specific parameters such as power,

data rate, coding etc. However, this real time adaptivity has an impact not only on

carrier-to-noise ratios but also on carrier-to-interference ratios and on upper layers.

13

System Description

Both aspects have therefore to be carefully studied. Various methods exist to counter-

act propagation effects at the physical layer level. The most relevant ones should take

into account operating frequency bands, performance objectives of the system and ge-

ometry of the network (system architecture, multiple access schemes).In fact FMT for

the physical layer can be divided into :

1. Power Control : transmitting power level fitted to propagation impairments,

2. Adaptive waveform : fade compensated by amore efficient modulation and cod-

ing scheme,

3. Diversity : fade avoided by the use of another less impaired link,

4. Layer 2 : coping with the temporal dynamics of the fade.

3.5.1 Power control

Four types of Power Control FMT can be considered : Up-Link Power Control (ULPC),

End-to-End Power Control (EEPC), Down-Link Power Control (DLPC) and On-Board

Beam Shaping (OBBS).Various ways of power control are explained in fig. 4.1

Up-Link Power Control (ULPC)

The aim of ULPC, the output power of a transmitting Earth station is matched to

uplink impairments. Transmitter power is increased to counteract fade or decreased

when more favorable propagation conditions are recovered so as to limit interference

in clear sky conditions and therefore to optimise satellite capacity. In the case of trans-

parent payloads, ULPC can prevent from reductions of satellite EIRP caused by the

decreased uplink power level that would occur in the absence of ULPC.

End-to-End Power Control(EEPC)

EEPC can be used for transparent configuration only. Indeed, the output power of a

transmitting Earth station is matched to up-link or down-link impairments. In the case

of regenerative repeaters, up and down links budgets are independent, so the concept

of EEPC can not exist anymore. EEPC is used to keep a constant overall margin of the

system. As for ULPC, transmitter power is increased to counteract fade or decreased

when more favourable propagation conditions are recovered to limit interference and

optimise satellite capacity.

14

System Description

Figure 3.2: Various ways of power control

Down-Link Power Control (DLPC)

WithDLPC, the on-board channel output power is adjusted to themagnitude of down-

link attenuation. DLPC aims to allocate a limited extra-power on-board in order to

compensate a possible degradation in term of down-link C/N0 due to propagation

conditions on a particular region. In this case, all Earth stations in the same spot beam

benefit from the improvement of EIRP.

On-Board Beam Shaping (OBBS)

OBBS technique is based on active antennas, which allows spot beam gains to be

adapted to propagation conditions. Actually, the objective is to radiate extra-power,

and to compensate rain attenuation only on spot beams where rain is likely to occur.

15

System Description

3.5.2 Adaptive waveform

These FMTs could be split into Adaptive Coding (AC), Adaptive Modulation (AM)

and Data Rate Reduction (DRR).

Adaptive Coding (AC)

The introduction of redundant bits to the information bits when a link is experiencing

fading, allows detection and correction of errors (FEC) caused by propagation impair-

ments and leads to a reduction of the required energy per information bit. Adaptive

coding consists in implementing a variable coding rate matched to impairments orig-

inating from propagation conditions.

Adaptive Modulation (AM)

Higher system capacity for a given bandwidth can be achieved with spectral efficient

modulation schemes but in clear sky conditions only due to link budget power limita-

tion. As Adaptive Coding, the aim of Adaptive Modulation is to decrease the required

energy per information bit required corresponding to a given BER, which translates

into a reduction of the spectral efficiency as C/N0 decreases. The reduction of the

spectral efficiency is the results of the use of lower level modulation schemes.

Data Rate Reduction (DRR)

Further reduction can be obtained by a decrease of the information data rate at con-

stant BER. The technique is called Data Rate Reduction. Here, user data rates should

be matched to propagation conditions : nominal data rates are used under clear sky

conditions (no degradation of the service quality with respect to the system margin),

whereas reductions is introduced according to fade levels.

3.5.3 Diversity

The objective of these techniques is to re-route information in the network in order

to avoid impairments due to an atmospheric perturbation. Three types of diversity

techniques can be considered: site (SD), satellite (SatD) and frequency (FD) diversity.

These techniques are very expensive as the associated equipments have to be redun-

dant.

Site Diversity(SD)

SD is based on the change of the network routes, therefore, it applies only for the Fixed

Satellite Service. SD takes advantage of the fact that two fades experienced by two

16

System Description

Figure 3.3: site diversity

Earth Stations separated by a distance higher than the size of a convective rain cell (at

least 10 km), are statistically independent. The Earth station affected by aweaker event

is used and the information is routed to the original destination through a separated

terrestrial network.The concept is explained with the help of fig. 4.2

Satellite Diversity(SatD)

Satellite Diversity can be regarded in two different ways : on one hand, when de-

signing the system, by optimizing the size of the constellation (that is the number of

satellites) in order to prevent communications at low elevation angles. On the other

hand in allowing Earth Stations to choose between various satellites, the one for which

the most favorable link with respect to the propagation conditions exists.

Frequency (FD) diversity

Frequency Diversity is a technique based on the fact that payloads using two different

frequency bands are available onboard. When a fade is occurring, links are re-routed

using the lowest frequency band payload, less sensitive to atmospheric propagation

impairments.

17

System Description

3.5.4 Layer 2

FMT at layer 2 level are techniques which do not aim at mitigating a fade event but

instead rely on the re-transmission of the message. Two different techniques can be en-

visaged at layer 2 : Automatic Repeat Request (ARQ) and Time Diversity (TD). With

ARQ, the message is sent regularly until the message reaches successfully the receiver.

ARQ with a random or predefined time repetition protocol would be an alternate so-

lution.

Time diversity can be considered as a FMT that aims to re-send the information when

the state of the propagation channel allows to get through. In this case, most often,

there is no need to receive the data file in real time and it is acceptable for the user

point of view to wait for the end of the propagation event (in general some tens of

minutes) or for a decrease of traffic. This technique benefits from the use of propa-

gation mid-term prediction model in order to estimate the most appropriate time to

re-sent the message without repeating the request.

First we will find the channel fade then we will take the decision whether we need to

activate the fade mitigation technique or not, if yes then we need to decide what is the

kind of fademitigation technique we need select. In this waywe need to adapt the link

according to channel conditions. Adaptive coding and modulation is the main mitiga-

tion technique that we are going to implement. Link adaptation is clearly explained in

the next section.

3.6 Link adaptation

As we all know that the link between between transmitter and receiver is wireless

in satellite communication and the future satellite communication is aiming to go for

higher frequency bands like Ka band. The use of the Ka band (30/20 GHz) for satel-

lite communication systems raises the problem of dealing with rain attenuation. As

opposed to the traditionally used Ku band (14/12 GHz), the Ka band is much more af-

fected by atmospheric events that lead to bad signal conditions, ranging from a slowly

changing attenuation of the signal to a sudden deep fade that blocks all communica-

tion. The link adaptation concept is taken from [7].

The channel fades can be tracked / predicted then transmission signal may be de-

signed so as to avoid the fades / take advantage of good channel conditions. Such

types of systems are known as link adaptation systems where link level parameters

are dynamically adjusted in order to maximize the data rates over a certain period of

time.

18

System Description

−8 −6 −4 −2 0 2 4 6 8 10 1210

−5

10−4

10−3

10−2

10−1

100

BER performance for M = 16modulation

Eb/No in dB

prob

. of e

rror

16−QAMAWGN

, 1/2 Conv.

16−QAMAWGN

16−QAMAWGN+RLY

16−QAMAWGN+RLY

, 1/2 Conv.

16−QAMAWGN

, 1/3 Conv.

16−QAMAWGN+RLY

, 1/3 Conv.

Figure 3.4: EbNo verses probability of error curves

3.6.1 BER performance for different modulation schemes

The bit error rate is used as the performance measure in satellite communication. The

bit error rate performance of different modulation schemes and coding rates are dif-

ferent. Some of the simulated curves are shown fig. 4.3. the simulation procedure is

as followed. first we need to generate the data bits then transmit these generated data

bits with some modulation scheme(eg. BPSK) with no code rate. At the receiver end

receive the bits and demodulate the bits. Compare the transmitted and received bits

and find out the probability of error(BER). Find out the BER for different values of

SNR. Repeat the same thing for different modulation schemes and coding rates and

plot the curves. Some of the simulated curves are shown fig. 4.3. From the curves we

can observe that as the SNR increases the probability of error decreases. In the fig. 4.3

the x-axis is EbNo and y -axis represents the corresponding probability of errors. As

we all know that as EbNo increases the corresponding BER will decreases. The differ-

ent curves are for different code rates for QAM and considered with AWGN channel.

When we calculate the link budget we will come to know the amount of margin we

have to operate. Suppose for successful operation of the link the minimum probability

of error required is 0.01, then from the fade margin we will select the suitable modu-

lation scheme. This kind of system is known as link adaptation system where system

parameters are changed according to the fade conditions.

If the EbNo is less then we will go for lower order modulation schemes but we need

19

System Description

Figure 3.5: SNR verses BER curves taken from [7]

to maintain the required probability of error. As the EbNo increases we can use the

higher order modulation schemes with different coding rates.

As an example consider the curves shown in fig. 4.4 taken from [7] in which the

BER performance of different modulation schemes and coding rates is given. If the

received SNR is below 8 dB none of the curves satisfy the required BER, hence it is

better not to transmit anything during that time.If the SNR is between 8 and 10 dB

better to transmit with QPSK because it satisfy the required BER. When it is between

10 and 15 dB both QPSK and 16-QAMwith c=1/3 satisfy the required BER but we will

use the higher order modulation scheme to send the data so that we can get the more

data rate. In this way the link is adaptively selected according to the SNR. Description

Mitigation Techniques

20

Chapter 4

Fade Mitigation Techniques

Satellite systems require radio spectrum. As the GEO orbit is filled up with satellites

operting in C and Ku bands the satellites were built for next higher frequency bands to

offer broader transmission channels for multimedia applications. An increasing num-

ber of new services are being promoted for Ka-band (20/30 GHz) satellite systems,

involving very small aperture terminals (VSAT). At the Ka-band,propagation impair-

ments strongly limit the quality and availability of satellite communications.

Inorder to avoid the fade we need to go for fade mitigation techniques. Before go-

ing into fade mitigation techniques, first of all we need to know the different kinds

of impairments that are there in the atmosphere. Then we need to find the amount

of attenuation presented by each impairment so that we can find the overall attenua-

tion presented by atmosphere. There are many phenomena that lead to signal loss on

transmission through the earths atmosphere like Atmospheric absorption, Cloud at-

tenuation, Ionospheric scintillation, Tropospheric scintillation, and Rain attenuation.

The detailed study of these attenuation models has been presented in ITU documents.

4.1 Propagation effects and their impact on satellite-earth

links

All radio wave signals have to be transmitted through the atmosphere. These signals

will be effected by the atmospheric impairments. The effects of atmosphere have to

be considered in system design at frequencies above 20 GHz. there are different kinds

of atmospheric impairments like atmospheric absorption, cloud attenuation, Tropo-

spheric scintillation, Low angle fading, Ionospheric scintillation and Rain attenuation.

A brief summary of atmospheric impairments will be presented in this section.

21


Atmospheric absorption

At microwave frequencies and above, electromagnetic waves interact with molecules

in the atmosphere to cause signal attenuation. At certain frequencies, resonant absorp-

tion occurs and severe attenuation can result.The amount of attenuation is less than 1

dB on most paths below 100 GHz.

Cloud attenuation

Clouds have become an important factor for someKa-band paths and all V-band(50/40

GHz) systems. The difficulty with modeling cloud attenuation is that clouds are of

many types and can exist at many levels.The water droplet concentrations in each

cloud will also vary,and clouds made up of ice crystals cause little attenuation. The

amount of attenuation is 1 and 2 dB at frequencies around 30 GHz

Tropospheric scintillation

Energy from the sun warms the surface of the earth and the resultant convective activ-

ity agitates the boundary layer. This agitation results in turbulent mixing of different

parts of boundary layer, causing small scale variations in refractive index. The rapid

variations in refractive index along the path will lead to fluctuations in the received

signal level these fluctuations are known as tropospheric scintillations.

Low angle fading

When the elevation angle falls below 10 degrees, a second propagation effect becomes

noticeable that is low angle fading. Low angle fading is the same phenomena as mul-

tipath fading in terrestrial paths. A signal transmitted from a satellite arrives at the

earth station receiving antenna via different paths with different phase shifts. On the

combination, the resultant waveformmay be enhanced or attenuated from the normal

clear sky level.

Ionospheric scintillation

Energy from sun causes the ionosphere to grow during the day, increasing the total

electron content(TEC) by two orders of magnitude, or more. The rapid change in

TEC from the daytime to nighttime, which occurs at local sunset in the ionosphere,

that gives rise to irregularities in the ionosphere. These rapid fluctuations are called

ionospheric scintillations.

22


Rain attenuation

At frequencies above 10 GHz , rain is the dominant propagation phenomenon on

satellite links.Rain drops absorb and scatter the electromagnetic waves. In Ku and

Ka bands rain attenuation is almost entirely caused by absorption.At Ka band there is

a small contribution from scattering by large rain drops.Rain is the primary cause of

depolarization. Atmospheric gases and tropospheric scintillation do not cause signal

depolarization.Ionosphere causes the depolarization.Some of the energy in one polar-

ization can cross over to other polarization due to asymmetric particles in the existing

path, which leads to depolarization. Measure of depolarization that is most useful

in analyzing communication system is the cross pole isolation(XPI).It is the ratio of

wanted power to the unwanted power.

23


4.1.1 Link Budget

Once we come to know the amount of attenuation in the atmosphere we will calculate

the link budget. Link budget is nothing but the tabular method of calculating the

received power and noise power.

Link Analysis for Regenerative Payload(Fair weather)

Link budget is nothing but the tabular method of calculating the received power and

noise power. When we calculate our link budget then we come to know how much

amount of margin we have and accordingly we will go for implementing the fade

mitigation techniques.

Parameters Unit NB Terminal WB TerminalTransmit Frequency GHz 29.6 to 30.2 29.6 to 30.2Receive Frequency GHz 20.6 to 21.6 20.6 to 21.6Antenna size Meter 0.28 offset feed parabolic 0.75 offset feed parabolicUP LINKSSPA Power output W 5.0 10.0Terminal Antenna gain dB 36.3 45.2Terminal EIRP dBW 42.8 54.35Satellite G/T dB/K 6.0 6.0C/No Available dBHz 62.9 74.4Data rate Kbps 64 2048C/No req for BER of 10−7 dBHz 57.1 72.1Available Link Margin dB 5.8 2.35DOWN LINK

SSPA EIRP dBW 40.0 42.0Terminal Antenna gain dB 33.2 41.76System noise temp K 315 315Terminal G/T dB/K 8.7 17.2C/No Available dBHz 66.8 77.3Data rate Kbps 576 2048C/No req for BER of 10−6 dBHz 63.6 69.1Available Link Margin dB 3.2 8.26

Table 4.1: Link Analysis for Regenerative Payload(Fair weather)

24


Link Analysis for Bent pipe Payload(Fair weather)

Link Analysis for Bent pipe Payload for Fair weather has been shown in the table.

Parameters Unit NB TerminalTransmit Frequency GHz 29.6 to 30.2Receive Frequency GHz 20.6 to 21.6Antenna size Meter 2.4 CassegrainUP LINKSSPA Power output W 13.0Terminal Antenna gain dB 55.21Terminal EIRP dBW 67.22Satellite G/T dB/K 6.0C/No Available dBHz 88.13DOWN LINKSSPA EIRP dBW 44.31Terminal Antenna gain dB 52.06Terminal G/T dB/K 27.56C/No Available dBHz 89.93Total C/No dBHz 86.33Data rate Mbps 40.00C/No req for BER of 10−6 dBHz 82.02Available Link Margin dB 4.31

Table 4.2: Link Analysis for Bent pipe Payload(Fair weather)

4.2 Fade Mitigation Techniques

As the operating frequency is increased, the atmospheric attenuations becomes more

severe. so implementing static margin as the only mean to compensate the propaga-

tion impairments at high frequency bands is not a good task, and it will push towards

the implementation of Fade Mitigation Techniques(FMT).Those techniques allow sys-

tems with rather small static margin to be designed, while overcoming in real time

cloud attenuation, some fraction of rain attenuation,scintillation, and depolarization

events.

Making use of FadeMitigation Techniques involves adapting in real time the link bud-

get to the propagation conditions through some specific parameters such as power,

data rate, coding etc. However, this real time adaptivity has an impact not only on

carrier-to-noise ratios but also on carrier-to-interference ratios and on upper layers.

Both aspects have therefore to be carefully studied. Various methods exist to counter-

act propagation effects at the physical layer level. The most relevant ones should take

into account operating frequency bands, performance objectives of the system and ge-

25


Figure 4.1: Various ways of power control

ometry of the network (system architecture, multiple access schemes).In fact FMT for

the physical layer can be divided into :

1. Power Control : transmitting power level fitted to propagation impairments,

2. Adaptive waveform : fade compensated by amore efficient modulation and cod-

ing scheme,

3. Diversity : fade avoided by the use of another less impaired link,

4. Layer 2 : coping with the temporal dynamics of the fade.

4.2.1 Power control

Four types of Power Control FMT can be considered : Up-Link Power Control (ULPC),

End-to-End Power Control (EEPC), Down-Link Power Control (DLPC) and On-Board

Beam Shaping (OBBS).Various ways of power control are explained in fig. 4.1

26


Up-Link Power Control (ULPC)

The aim of ULPC, the output power of a transmitting Earth station is matched to

uplink impairments. Transmitter power is increased to counteract fade or decreased

when more favorable propagation conditions are recovered so as to limit interference

in clear sky conditions and therefore to optimise satellite capacity. In the case of trans-

parent payloads, ULPC can prevent from reductions of satellite EIRP caused by the

decreased uplink power level that would occur in the absence of ULPC.

End-to-End Power Control(EEPC)

EEPC can be used for transparent configuration only. Indeed, the output power of a

transmitting Earth station is matched to up-link or down-link impairments. In the case

of regenerative repeaters, up and down links budgets are independent, so the concept

of EEPC can not exist anymore. EEPC is used to keep a constant overall margin of the

system. As for ULPC, transmitter power is increased to counteract fade or decreased

when more favourable propagation conditions are recovered to limit interference and

optimise satellite capacity.

Down-Link Power Control (DLPC)

WithDLPC, the on-board channel output power is adjusted to themagnitude of down-

link attenuation. DLPC aims to allocate a limited extra-power on-board in order to

compensate a possible degradation in term of down-link C/N0 due to propagation

conditions on a particular region. In this case, all Earth stations in the same spot beam

benefit from the improvement of EIRP.

On-Board Beam Shaping (OBBS)

OBBS technique is based on active antennas, which allows spot beam gains to be

adapted to propagation conditions. Actually, the objective is to radiate extra-power,

and to compensate rain attenuation only on spot beams where rain is likely to occur.

4.2.2 Adaptive waveform

These FMTs could be split into Adaptive Coding (AC), Adaptive Modulation (AM)

and Data Rate Reduction (DRR).

Adaptive Coding (AC)

The introduction of redundant bits to the information bits when a link is experiencing

fading, allows detection and correction of errors (FEC) caused by propagation impair-

27


ments and leads to a reduction of the required energy per information bit. Adaptive

coding consists in implementing a variable coding rate matched to impairments orig-

inating from propagation conditions.

Adaptive Modulation (AM)

Higher system capacity for a given bandwidth can be achieved with spectral efficient

modulation schemes but in clear sky conditions only due to link budget power limita-

tion. As Adaptive Coding, the aim of Adaptive Modulation is to decrease the required

energy per information bit required corresponding to a given BER, which translates

into a reduction of the spectral efficiency as C/N0 decreases. The reduction of the

spectral efficiency is the results of the use of lower level modulation schemes.

Data Rate Reduction (DRR)

Further reduction can be obtained by a decrease of the information data rate at con-

stant BER. The technique is called Data Rate Reduction. Here, user data rates should

be matched to propagation conditions : nominal data rates are used under clear sky

conditions (no degradation of the service quality with respect to the system margin),

whereas reductions is introduced according to fade levels.

4.2.3 Diversity

The objective of these techniques is to re-route information in the network in order

to avoid impairments due to an atmospheric perturbation. Three types of diversity

techniques can be considered: site (SD), satellite (SatD) and frequency (FD) diversity.

These techniques are very expensive as the associated equipments have to be redun-

dant.

Site Diversity(SD)

SD is based on the change of the network routes, therefore, it applies only for the Fixed

Satellite Service. SD takes advantage of the fact that two fades experienced by two

Earth Stations separated by a distance higher than the size of a convective rain cell (at

least 10 km), are statistically independent. The Earth station affected by aweaker event

is used and the information is routed to the original destination through a separated

terrestrial network.The concept is explained with the help of fig. 4.2

28


Figure 4.2: site diversity

Satellite Diversity(SatD)

Satellite Diversity can be regarded in two different ways : on one hand, when de-

signing the system, by optimizing the size of the constellation (that is the number of

satellites) in order to prevent communications at low elevation angles. On the other

hand in allowing Earth Stations to choose between various satellites, the one for which

the most favorable link with respect to the propagation conditions exists.

Frequency (FD) diversity

Frequency Diversity is a technique based on the fact that payloads using two different

frequency bands are available onboard. When a fade is occurring, links are re-routed

using the lowest frequency band payload, less sensitive to atmospheric propagation

impairments.

4.2.4 Layer 2

FMT at layer 2 level are techniques which do not aim at mitigating a fade event but

instead rely on the re-transmission of the message. Two different techniques can be

envisaged at layer 2 : Automatic Repeat Request (ARQ) and Time Diversity (TD).

29


With ARQ, the message is sent regularly until the message reaches successfully the

receiver. ARQ with a random or predefined time repetition protocol would be an

alternate solution.

Time diversity can be considered as a FMT that aims to re-send the information when

the state of the propagation channel allows to get through. In this case, most often,

there is no need to receive the data file in real time and it is acceptable for the user point

of view to wait for the end of the propagation event (in general some tens of minutes)

or for a decrease of traffic. This technique benefits from the use of propagation mid-

term prediction model in order to estimate the most appropriate time to re-sent the

message without repeating the request.

First we will find the margin available then we will take the decision whether we

need to activate the fade mitigation technique or not, if yes then we need to decide

what is the kind of fade mitigation technique we need select. in this way we need

to adapt the link according to channel conditions. Adaptive coding and modulation

is the main mitigation technique that we are going to implement. Link adaptation is

clearly explained in the next section.

4.3 Link adaptation

As we all know that the link between between transmitter and receiver is wireless

in satellite communication and the future satellite communication is aiming to go for

higher frequency bands like Ka band. The use of the Ka band (30/20 GHz) for satel-

lite communication systems raises the problem of dealing with rain attenuation. As

opposed to the traditionally used Ku band (14/12 GHz), the Ka band is much more af-

fected by atmospheric events that lead to bad signal conditions, ranging from a slowly

changing attenuation of the signal to a sudden deep fade that blocks all communica-

tion.

The channel fades can be tracked / predicted then transmission signal may be de-

signed so as to avoid the fades / take advantage of good channel conditions. Such

types of systems are known as link adaptation systems where link level parameters

are dynamically adjusted in order to maximize the data rates over a certain period of

time.




ferent. Some of the simulated curves are shown fig. 4.3. the simulation procedure is

30


−8 −6 −4 −2 0 2 4 6 8 10 1210

−5

10−4

10−3

10−2

10−1

100

BER performance for M = 16modulation

Eb/No in dB

prob

. of e

rror

16−QAMAWGN

, 1/2 Conv.

16−QAMAWGN

16−QAMAWGN+RLY

16−QAMAWGN+RLY

, 1/2 Conv.

16−QAMAWGN

, 1/3 Conv.

16−QAMAWGN+RLY

, 1/3 Conv.

Figure 4.3: EbNo verses probability of error curves

as followed. first we need to generate the data bits then transmit these generated data

bits with some modulation scheme(eg. BPSK) with no code rate. At the receiver end

recieve the bits and demodulate the bits. Compare the transmitted and received bits

and find out the probability of error(BER). Find out the BER for different values of

SNR. Repeat the same thing for different modulation schemes and coding rates and

plot the curves. Some of the simulated curves are shown fig. 4.3. From the curves we

can observe that as the SNR increases the probability of error decreases. In the fig. 4.3

the x-axis is EbNo and y -axis represents the corresponding probability of errors. As

we all know that as EbNo increases the corresponding BER will decreases. The differ-

ent curves are for different code rates for QAM and considered with AWGN channel.

When we calculate the link budget we will come to know the amount of margin we

have to operate. Suppose for successful operation of the link the minimum probability

of error required is 0.01, then from the fade margin we will select the suitable modu-

lation scheme. This kind of system is known as link adaptation system where system

parameters are changed according to the fade conditions.

If the EbNo is less then we will go for lower order modulation schemes but we need

to maintain the required probability of error. As the EbNo increases we can use the

higher order modulation schemes with different coding rates.

As an example consider the curves shown in fig. 4.4 taken from [7] in which the

BER performance of different modulation schemes and coding rates is given. If the

received SNR is below 8 dB none of the curves satisfy the required BER, hence it is

31


Figure 4.4: SNR verses BER curves taken from [7]

better not to transmit anything during that time.If the SNR is between 8 and 10 dB

better to transmit with QPSK because it satisfy the required BER. When it is between

10 and 15 dB both QPSK and 16-QAMwith c=1/3 satisfy the required BER but we will

use the higher order modulation scheme to send the data so that we can get the more

data rate. In this way the link is adaptively selected according to the SNR. Mitigation

Techniques of FMT

32

Chapter 5

Implementation of FMT

5.1 FMT control logic

The aim of a FMT control loop is to track the variations of the propagation channel in

real time and to compensate propagation impairments either to increase its availabil-

ity or to improve its instantaneous performance. For this purpose, it is first necessary

to detect when a fade is occurring in order to assess if the quality of link is going to

be degraded or if an outage is going to occur. Secondly, whenever an event supposed

to lead to an outage is detected, it is necessary to check if the terminal is authorized

to set up the mitigation, and upon reception of the clearance, to trigger the mitigation

process. Another step can consist in performing a real time prediction of the propaga-

tion channel in order to compensate the reaction time of the system to obtain a better

control loop behaviour.The system block diagram is explained with the help of fig. 5.1

where the transmitter will send the beacon signal.The receiver gets the transmitted

signal and sends back the channel quality through the feedback.accordingly the link

parameters will be changed to counteract the fade.The basic FMT logic is explained

with the help of fig. 8.1

5.2 Implementation of FMT

The electromagnetic waves undergo power decrease, scattering and depolarization

during propagation through rain and clouds. The attenuation due to rain is the main

factor that influences the performance of a high-frequency satellite communications

link. This results in a decrease of the percentage of the time for which a satellite link

can be expected to operate with a specified bit error rate (BER). Consequently, atmo-

spheric impairments affect availability and throughput of the communication system.

In order to overcome the adverse effects of atmosphere and to improve the reliability

of communication system we should control the parameters of system. This can be

33


Figure 5.1: System block diagram

Figure 5.2: Block diagram of FCM

accomplished by a fade countermeasure system. The block diagram of a fade counter-

measure system is shown in fig. 5.2

In the block diagram transmitting part constitutes the error control coding and bit to

symbol mapping. These transmitted signal is multiplied by the channel coefficients

and added by the noise. The receiving part will detect the faded symbols and de-

modulate to get the original bit stream. The knowledge of channel conditions can be

34


obtained by observing the primary parameters like attenuation, statistical distribution

of rain or secondary parameters like signal to noise ratio, BER, received signal char-

acteristics. Once we know the amount of fading present in the channel(by detection)

at time ’T’ we will try to predict the channel information at T+t. If we can predict the

channel information well in advance we can make a decision to adapt the coding and

modulation according to channel information.

5.3 Description of simulator

5.3.1 CRC Encoding

The data-bits are padded with the CRC bits. The CRC polynomial used for this pur-

pose is 0x04C11DB7, which is the IEEE 802.3 standard for Ethernet links. The basic

idea is to take each packet of data bits at a time, add 32-bit CRC code to it and give this

modified packet to the modulator to be sent across the link.

5.3.2 Error Control Coding

The channel coding is used for the reliable transmission of digital information over

the channel. The error control coding techniques rely on the systematic addition of re-

dundant symbols to the transmitted information so as to facilitate two basic objectives

at the receiver i.e. error detection and error correction. The channel encoder accepts

message bits and adds redundancy according to a prescribed rule, thereby producing

encoded data at a higher bit rate. The channel decoder exploits the redundancy to

decide which message bit was actually transmitted. The channel goal of the channel

encoder and decoder is to minimize the effect of channel noise. That is, the number

of errors between the channel encoder input and the channel decoder output is mini-

mized.

5.3.3 Modulation

Modulation is defined as the process by which some characteristics of a carrier is var-

ied accordance with a modulating wave. In digital communications, the modulating

wave consists of binary data or an M-ary encoded version of it. There are different

types of modulation schemes available but the final priority is determined by the way

in which the available primary communication resources, transmitted power and the

channel bandwidth, are best exploited. The raw bits are converted into symbols and

the number of bits in each symbol varies according to the type of modulation used. If

35


0 200 400 600 800 1000 1200 1400 1600 18003.5

4

4.5

5

5.5

6

6.5

7

Time in sec

SN

R

Channel SNR

Figure 5.3: variation of channel with time

the modulation scheme is BPSK then we will have one bit per symbol and for QPSK it

will be two bits per symbol and for 16 QAM it will be four bits per symbol.

5.3.4 Channel

The transmitted signal(bits) is to be transmitted through the channel so modeling of

the channel has to be done. channel will attenuate the signal. The transmitted signal

is multiplied by the channel coefficient. The small fluctuations in the channel will lead

to fading. Hence to overcome this fading we need to implement the fade mitigation

techniques. Here we have got the received signal strength for some days. The data has

been collected for every 16 seconds that is the channel coefficient is changing for every

sixteen seconds. The variations of the channel with respect to time for one particular

day has been plotted.

5.3.5 Automatic Gain Control (AGC)

Automatic gain control (AGC) is an adaptive system found inmany electronic devices.

The average output signal level is fed back to adjust the gain to an appropriate level

for a range of input signal levels. For example, without AGC the sound emitted from

an AM radio receiver would vary to an extreme extent from a weak to a strong signal;

the AGC effectively reduces the volume if the signal is strong and raises it when it is

weaker. AGC algorithms often use a PID controller where the P term is driven by the

error between expected and actual output amplitude.In the simulation, AGC is not

36


being used for the time being.

5.3.6 demodulation

The received symbols which are in the form of constellation are converted to bits by

symbol de-mapping. Hence, the bits are extracted from the symbols.

5.3.7 Decoding

In this block, the received bits are decoded .The decoding techniques rely on the sys-

tematic removal of redundant symbols that were padded to the transmitted bits so as

to facilitate two basic objectives at the receiver error detection and error correction.

5.3.8 CRC Decoder

At the receiver end, the received bits at the end of the demodulator block are taken one

packet at a time, the CRC code added at the transmitter is stripped off and the packet

is checked for errors. The total number of erroneous packets is, thus, kept track of.

The channel,detection and the decision algorithms are described in the subsequent

chapters. of FMT

37

Chapter 6

Channel

To study the characteristics of the channel is the most primary requirement for model-

ing the same. Study of the existing literature has not yet led to a strong conclusion in

this matter. Excerpts from some of the references follow:

6.1 Literature review

Reference [24]: The model describes that the rain attenuation probability distribution

is predicted dependent on two sample values measured shortly earlier. The model

equations and its parameter values have been derived empirically from measurement

results. For a signal with a sampling time of 10 seconds, the probability distribution

function of the attenuation of any sample is predicted as the ’hyperbolic secant distri-

bution’. This model shows the spectral characteristics of short-term dynamic during

a rain attenuation event are well described. The synthesized time series exhibit a de-

creasing power spectral density when the event duration increases, which is caused

by the procedure to obtain the desired maximum event attenuation is shown in this

paper.

Reference [25]: This paper has described the standard deviation of attenuation as the

function of the previous two value of attennuation. Once the average and standard

deviation of next rain attenuation is known, its probability distribution function can

be plotted a short time after a measured value.

Reference [26]: This model provides the relationships between the parameters .The

dynamic model is based on the log normal distributions of the rain attenuation and

utilizes a non-linear device to transform attenuation and rain intensity into a one-

dimensional Gaussian Stationary Markov process. So, the rate of change of attenua-

tion, mean value of attenuation and the variance are taken care of in the simulation.

During transmission through the channel, the signal gets attenuated due to the signal

amplitude fluctuation that leads to fading. To overcome the problem, fade mitiga-

38

Channel

tion techniques need to be implemented. As the first step to proceed in the domain

of channel modeling, the measured data regarding the received signal strength were

collected for a few days. The data has been collected for every 10 seconds which is

also considered as the sampling time in the simulation. That is, the channel coefficient

is changing every ten seconds.

6.2 Channel considered in simulation

In the simulation, the channel is modeled as a slow-varying Rayleigh channel with

the negligible Doppler effect. The output of the channel is taken to be the attenuation

values in the simulation. The noise power also varies according to these attenuation

values, due to coupling of the increased sky noise due to rain into the receiver antenna.

The noise power Pn in watts [4] in this case is a function of attenuation. The noise

power is given by

Pn = KTsBn (6.1)

where K = The Boltzmann’s constant = -228.6 dBW/K/Hz

Ts = The physical temperature of the source in Kelvin degrees

Bn = The noise bandwidth in which the power is measured in Hertz

The total path attenuation A(dB) is the sum of the clear sky attenuation due to atmo-

spheric gaseous absorption A(clear air) and the attenuation due to the rain,A(rain)

A = Aclearair + AraindB (6.2)

The sky noise temperature T(sky) resulting from the total path attenuation A (dB) is:

Tsky = 270(1− 10−A/10)K (6.3)

where 270 K is the assumed temperature of the medium due to rain [9].

The antenna noise temperature T(antenna) is calculated by multiplying the coupling

coefficient (n of 90-95 percent) with T(sky). Thus,

Tantenna = nTskyK (6.4)

As the satellites use a high gain LNA, the contribution of the later parts of the receiver

to the system noise temperature becomes negligible. The system noise temperature

T(system) is:

Tsystem = TLNA + TantennaK (6.5)

From the above expressions, it is found that the temperature varies with the atten-

uation. As the attenuation increases the noise temperature increases that in subse-

39

Channel

quently increases the noise power. This concept has been incorporated in the simula-

tion, where with the change in attenuation, the noise power is varying.

40

Chapter 7

Detection

The FMT control loop will constitute three parts one is detection of impairments sec-

ond is prediction of impairments a short time ahead and finally it has to take a decision

whether to activate FMT or not. The amount of attenuation present in the channel is

measured through the measurements. Measurements is nothing but the detection of

attenuation. The objective of the detection function is to quantify the magnitude of a

fade event occurring on the considered link. Three kinds of detection concepts are

1. Open loop detection

2. Closed loop detection

3. Hybrid loop detection

Open-loop Detection

The open-loop detection concept relies on the estimation of uplink (or downlink) im-

pairment from a measurement of the propagation conditions. This measurement can

be carried out in several ways: rain intensity and other meteorological measurements,

sky brightness temperatures measured with a radiometer, radar networks, satellite

imagery or satellite beacon operating at uplink or downlink frequency.

Closed-loop Detection

In the closed-loop detection concept, estimation of the impairment is performed from

the measurement of the overall link performance. Bit Error Rate or Carrier plus Noise

estimations can be carried out by the earth station [6] or by the satellite (if On-Board-

Processing enables so). In the case of a transparent satellite link, a measurement of the

overall link will give information on the total degradation of the propagation channel.

However, it will not identify if the impairment is occurring on the uplink or on the

downlink.

41

Detection

Hybrid-loop Detection

To separate uplink and downlink fade contributions the hybrid-loop detection concept

uses two different measurements, one of them from a beacon and the other from the

link .

7.1 Methods from the literature

There are many link quality estimation algorithms available in the literature such as:

BER Counter Method Pseudo-error Method Method of Mean Method of Extended

Quantization However, all the above methods except the first one need access to the

internal components of the satellite modem. But we would not have access to the dif-

ferent demodulator blocks in our case as we would be using an off-the-shelf modem

provided by ISRO. Thus, we decided to use the BER counter method. A brief discus-

sion of the method is given here.

7.2 Fade detection

7.2.1 Fade detection using CRC

Estimation of the SNR is the first step in the operation of a FMT system. We propose

to estimate the SNR using the PER. Since the modulation and code-rate is known be-

forehand while the link is in operation, an estimate of the PER will allow us to get an

estimate of the current SNR. Using this value of the SNR, we can go on to decide the

optimum ACM scheme to use in the particular situation. We propose to estimate the

SNR using CRC32 to detect packet errors.

A cyclic redundancy check (CRC) or polynomial code checksum is a non-secure hash

function designed to detect accidental changes to raw computer data, and is com-

monly used in digital networks and storage devices such as hard disk drives. A CRC-

enabled device calculates a short, fixed-length binary sequence, known as the CRC

code or just CRC, for each block of data and sends or stores them both together. When

a block is read or received the device repeats the calculation; if the new CRC does not

match the one calculated earlier, then the block contains a data error and the device

may take corrective action such as rereading or requesting the block be sent again,

otherwise the data is assumed to be error free.

CRCs are so called because the check (data verification) code is a redundancy (it adds

zero information) and the algorithm is based on cyclic codes. The term CRC may re-

42

Detection

Figure 7.1: SNR detection using CRC method

fer to the check code or to the function that calculates it, which accepts data streams of

any length as input but always outputs a fixed-length code. CRCs are popular because

they are simple to implement in binary hardware, are easy to analyse mathematically,

and are particularly good at detecting common errors caused by noise in transmis-

sion channels. The CRC was invented by W. Wesley Peterson, and published in his

1961 paper. The earliest known appearance of the 32-bit polynomial most commonly

used by standards bodies was in a 1975 paper written by the Georgia Institute of Tech-

nology on behalf of the Rome Laboratory. The CRC is an error-detecting code. Its

computation resembles a polynomial long division operation in which the quotient is

discarded and the remainder becomes the result, with the important distinction that

the polynomial coefficients are calculated according to the carry-less arithmetic of a

finite field. The length of the remainder is always less than the length of the divisor

(called the generator polynomial), which therefore determines how long the result can

be.

Although CRCs can be constructed using any finite field, all commonly used CRCs

employ the finite field GF(2). This is the field of two elements, usually called 0 and 1,

comfortably matching computer architecture. The rest of this article will discuss only

these binary CRCs, but the principles are more general.

An important reason for the popularity of CRCs for detecting the accidental alteration

of data is their efficiency guarantee. Typically, an n-bit CRC, applied to a data block

of arbitrary length, will detect any single error burst not longer than n bits (in other

words, any single alteration that spans no more than n bits of the data), and will detect

a fraction 1− 2−n of all longer error bursts. Errors in both data transmission channels

and magnetic storage media tend to be distributed non-randomly (i.e. are ”bursty”),

making CRCs’ properties more useful than alternative schemes such as multiple par-

ity checks. The simplest error-detection system, the parity bit, is in fact a trivial CRC:

it uses the two bit long divisor “11”.

The CRC polynomial used by us is 0x04C11DB7, which is the IEEE 802.3 standard for

Ethernet links. The basic idea is to take each packet of data bits at a time, add 32-bit

43

Detection

CRC code to it and give this modified packet to the modulator to be sent across the

link. At the receiver end, the received bits at the end of the demodulator block are

taken one packet at a time, the CRC code added at the transmitter is stripped off and

the packet is checked for errors. The total number of erroneous packets is, thus, kept

track of.

We have implemented the SNR estimation in the simulator in a similar fashion. The

built-in MATLABr functions generate() and detect() have been used for this

purpose. The generator and detector objects required for the function to work are

created before starting the simulation. For each channel value, the data bits to be sent

are passed to the generate() function in the form of a column vector and the func-

tion adds the CRC code at the end of this vector. These bits are then sent modulated

and noise of appropriate power is added to model the channel. After being received at

the demodulator and then, decoded, the data bits are passed to the detect() function

which returns the data bits sans the CRC bits and a variable which indicates whether

or not the packet contained an error. The number of erroneous packets received is

counted using an accumulator variable.

7.2.2 Detection using Embedded pilot

The next method for SNR estimation is Embedded pilot. In this method, a sequence of

known pilot bits are transmitted through the channel. These pilot bits are received at

the receiver and the channel SNR value is estimated from the BER. Two variations of

this method are possible: Continuous pilot and Distributed pilot. The two approaches

are discussed below.

7.2.3 Continuous pilot

In the Continuous pilotmethod, the pilot bits are transmitted together. In other words,

instead of appending pilot bits to each and every packet, we send them together as

a single long sequence after sending a specified number of raw data packets. This

approach is very similar to that used in mobile communication, where a frame of data

sent consists of raw data

7.2.4 Detection using Embedded pilot

The next method for SNR estimation is Embedded pilot. In this method, a sequence of

known pilot bits are transmitted through the channel. These pilot bits are received at

the receiver and the channel SNR value is estimated from the BER. Two variations of

44

Detection

Figure 7.2: SNR detection using Continuous Pilot method

this method are possible: Continuous pilot and Distributed pilot. The two approaches

are discussed below.

Continuous pilot

In the Continuous pilot method, the pilot bits are transmitted together. In other words,

instead of appending pilot bits to each and every packet, we send them together as

a single long sequence after sending a specified number of raw data packets. This

approach is very similar to that used in mobile communication, where a frame of data

sent consists of raw data

An important point requiring attention is that the PER found out in this manner

is used to estimate the SNR and decide the optimum ACM scheme only in the next

iteration of the simulation. This has been done to model the delay typically associated

with earth-satellite links which are generally of the order of 500 ms. In our simulation,

we have assumed that the channel varies considerably only after 10 seconds. So, data

corresponding to 10 seconds is processed during each iteration of the main loop and

the total number of packet errors during these 10 seconds is calculated. This PER is

used to estimate the SNR and decide the optimum ACM scheme in the next iteration.

Thus, a channel-information feedback delay of 10 seconds is automatically included

in the simulation. The system can be studied for various values of feedback-delay in

the future.

7.3 Results




ferent. Some of the curves are shown fig. 8.3 In the fig. 8.3 the x-axis is SNR and y

-axis represents the corresponding probability of errors. As we all know that as SNR

45

Detection

increases the corresponding BER will decreases. The different curves are for different

code rates for QAM, QPSK and BPSK and considered with AWGN channel.

−2 0 2 4 6 8 10 12 14 16 18

10−4

10−3

10−2

10−1

BER performance for 4−QAM(similar to QPSK), 16−QAM and 64−QAM

SNR (in dB)

BE

R

BPSK simulatedBPSK theoreticalQPSK simulatedQPSK theoretical16−QAM simulated16−QAM theoreticalBPSK 1/2 rate codeBPSK 1/3 rate codeQPSK 1/2 rate codeQPSK 1/3 rate code16−QAM 1/2 rate code16−QAM 1/3 rate code

Figure 7.3: SNR verses probability of error curves

7.3.2 Performance of the system for the collected data

The performance of the system for the collected data has been presented below. We

have simulated the system for the collected data that is the rain attenuation data has

been collected for several days and experiments have been performed on these data.

Basically the results have been presented for two days one is on 5th where we have

huge amount of rain and second is on sixth wherewe do not have any rain attenuation.

Change of modulation scheme and coding rate with respect to time

Based on the value of channel conditions the modulation scheme and coding rate

should be changed. If the channel condition is poor we will go for lower order mod-

ulation schemes and if the channel condition is good then we will go for higher order

modulation schemes.

Actually the detection part has to detect the amount of fade present in the atmosphere

but we have using the rain data for simulation. The measurement of beacon signal

have been given that is the signal strength value is given based on which we need to

take the decision. Based on the values of signal strength we have selected the modula-

tion scheme and coding. In the figure. 8.4 we can observe that as the channel condition

is changing the modulation and coding rate are also changing. The bit error rate for

46

Detection

0 200 400 600 800 1000 1200 1400 1600 18000

2

4

6

8

10

12

14

16

Time in sec

SN

R

Channel SNRM values chosenC values chosen

Figure 7.4: Change of modulation and coding with time (date 6th)

0 20 40 60 80 100 120 140 160 180−20

−15

−10

−5

0

5

10

15

20

Time in sec

SN

R


Figure 7.5: Change of modulation and coding with time (date 5th)

different modulation schemes with respect to time is presented in the figure. 8.7.

In figure. 8.5 we can observe that during 60 -80 sec there is a heavy rain because of

which the signal strength goes down which in turn increments the BER. By using

ACM we can send the data but even then we will not have the enough margin to

overcome this outage. During that time we can go for some other FMTs like power

control in which we will increase the transmit power or site diversity in which link is

made available through some other terrestrial link. Joint FMTs will further improve

the system performance.

47

Detection

0 20 40 60 80 100 120 140 160 18010

−4

10−3

10−2

10−1

100

Time in sec

BE

R

M=2,C=1M=4,C=1M=16,C=1M=16, C=2BER, AMC

Figure 7.6: Change of BER with time using collected data (date 5th)

0 200 400 600 800 1000 1200 1400 1600 180010

−4

10−3

10−2

10−1

100

Time in sec

BE

R

M=4,C=1M=2,C=1M=16,C=1M=16,C=2BER, AMCThreshold

Figure 7.7: Change of BER with time using collected data (date 6th)

BER performance for the collected data

As the time varies the channel condition changes hence the bit error rate will also

changes. The change in BER at different point of times is shown in fig. 8.7. The thresh-

old BER that we have set was 10−1 and when we observe the simulated BER most of

the times it is below the target threshold if we use adaptive modulation and coding.

In the result we have shown the bit error rate for different modulation schemes. If we

48

Detection

use 16 QAM with 1/2 code rate we will achieve higher bit rate but we can observe

from the result that most of the times its bit error rate is above the threshold. Hence

if we use this constant rate transmission scheme then we will have outage most of the

times, which is not desirable. Same is the case with 16 QAM with rate 1/3. If we use

QPSK with code 1/2 and QPSK with no code then we can achieve the required bit

error rate but the data rate for these schemes will be less. So to achieve the higher data

rate and the required bit error rate we will go for adaptive coding and modulation,

that is based on the channel conditions we will change our modulation scheme and

coding rate. If the channel condition is poor then we will use lower order modulation

schemes and if the channel condition is good we will go for higher order modulation

schemes. the result is shown as BER, AMC in fig. 8.7. Hence we can conclude that by

using adaptive modulation and coding rate we can improve the data rate providing

the required bit error rate.

7.3.3 PER performance for different SNR values

The performance of the system with respect to packet error rate has been done here.

The packet size is taken as 1 Kb. With respect to the figure 8.8, it is seen that, when the

SNR is high, the per is less and the higher order modulation schemes and coding rates

are chosen. But, when the channel condition is worse, the lesser modulation schemes

and coding rates are applied so as to send as much less number of bits as possible.

0 5 10 15 20

10−1

100

PER performance for different ACM schemes

SNR (in dB)

PE

R

QPSK−1/2 simulatedQPSK−1/3 simulatedQPSK theoretical16−QAM simulated16−QAM−1/2 simulated16−QAM−1/3 simulated16−QAM theoretical64−QAM simulated64−QAM−1/2 simulated64−QAM−1/3 simulated64−QAM theoretical16−QAM theoretical16−QAM−1/2 simulated16−QAM−1/3 simulated64−QAM−1/2 simulated64−QAM−1/3 simulated

Figure 7.8: PER versus SNR curves

49

Detection

7.3.4 Performance of the FMT system with time

The performance of the system with the proposed detection algorithm has been pre-

sented below. The estimated SNR almost tracks the calculated SNR most of the time.

Based on the estimated SNR, the decision of adaptive modulation and coding scheme

is taken. The corresponding diagram is shown in 8.9

Figure 7.9: Time versus SNR and change of M and C

7.3.5 SNR estimation accuracy with No Back Off


sented below. The estimated SNR almost tracks the calculated SNRmost of the time in

case of CRC algorithm. Performance of Distributed Pilot and Continuous Pilot meth-

ods is almost similar, the former being a little better. Based on the estimated SNR, the

decision of adaptive modulation and coding scheme is taken.

7.3.6 SNR estimation accuracy with Symmetric Back Off






50

Detection

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit in sec

SN

R IN

dB

Comparision of SNR curves with NBF

SNR calculatedSNR estimated with NBF for CONT pilotSNR estimated with NBF for DIST pilot

Figure 7.10: SNR estimation accuracy betweenDistributed Pilot and Continuous Pilot methods

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit in sec

SN

R IN

dB


SNR calculatedSNR estimated with NBF for CONT pilotSNR estimated with NBF for CRC

Figure 7.11: SNR estimation accuracy between Continuous Pilot and CRC methods

7.3.7 SNR estimation accuracy with Asymmetric Back Off






51

Detection

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit in sec

SN

R IN

dB


SNR calculatedSNR estimated with NBF for CRCSNR estimated with NBF for DIST pilot

Figure 7.12: SNR estimation accuracy between Distributed Pilot and CRC methods

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit in sec

SN

R in

dB

Comparision of SNR curves with SBF

SNR calculatedSNR estimated with SBF for CONT pilotSNR estimated with SBF for DIST pilotSNR estimated with SBF for CRC

Figure 7.13: SNR estimation accuracy between Distributed Pilot, Continuous Pilot and CRCmethods

7.3.8 SNR estimation accuracy with Adaptive Back Off






52

Detection

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit insec

SN

R in

dB

Comparision of SNR curves with ASBF

SNR calculatedSNR estimated with ASBF for CONT pilotSNR estimated with ASBF for CRCSNR estimated with ASBF for DIST pilot

Figure 7.14: SNR estimation accuracy between Distributed Pilot, Continuous Pilot and CRCmethods

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit in sec

SN

R in

dB

Comparision of SNR curves with Adaptive back off

SNR calculatedSNR estimated with adaptive back off for DIST pilotSNR estimated with adaptive back off for CONT pilot

Figure 7.15: SNR estimation accuracy betweenDistributed Pilot and Continuous Pilot methods

53

Detection

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit in sec

SN

R in

dB


SNR calculatedSNR estimated with adaptive back off for CONT pilotSNR estimated with adaptive back off for CRC

Figure 7.16: SNR estimation accuracy between Continuous Pilot and CRC methods

0 50 100 150 200 250 300−2

0

2

4

6

8

10

12

time unit in sec

SN

R in

dB


SNR calculatedSNR estimated with adaptive back off for DIST pilotSNR estimated with adaptive BACKOFF for CRC

Figure 7.17: SNR estimation accuracy between Distributed Pilot and CRC methods

54

Chapter 8

Decision

Implementation of fade mitigation techniques will constitute three parts one is detec-

tion of the attenuation present in the channel at time T, second is prediction of the

channel coefficient at time T+t based on the detected value at T, third is the decision in

which one need to take the decision whether we need to activate the fade mitigation

technique or not and also the level of fade mitigation technique is also to be decided.

The FMT control logic is given in fig. 8.1. The channel modeling is given below and

the detection part has been given in the appendix. The decision part will be explained

in this chapter

8.1 Decision

The objective of decision function is to take a decision whether to activate FMT or not.

If FMT is to be activated what kind of FMT is to be considered. These things mainly

depends on the channel measurements. The idea is to implement the adaptive coding

and modulation. Using the information of the predicted attenuation, this function will

Figure 8.1: FMT control logic

55

Decision

decide if the considered link performs according to specifications that is

Eb

N0

≥Eb

N0

|Required (8.1)

for a given BER, modulation and coding scheme. Where

Eb

N0

=C

N0

⋆1

Rb

(8.2)

Where Rb is the information data rate. the above expression is equivalent to

Systemmargin ≥ 0 (8.3)

where

Systemmargin ≡Eb

N0

≥Eb

N0

|Required (8.4)

for a given modulation and coding scheme.

In clear sky conditions, equation (5.3) will be met. However during a fade event the

system margin may become negative and therefore a Fade Mitigation Technique shall

be activated. This technique will vary one of the clear sky systems parameters trans-

mitted power, antenna Gain, data rate, coding etc. Here the modulation scheme and

coding rate of the transmitter will be changed according to conditions.

8.1.1 Detection margin and Hysteresis

When attenuation is present in the link, equation 5.3 will determine whether the acti-

vation of the FMT is needed. However, to cope with possible estimation and predic-

tion errors, an additional margin can be included in the link budget equation

Systemmargin ≥ Controllogicmargin (8.5)

When system margin goes below the Control logic Margin Threshold, the Fade Miti-

gation Technique will be activated by changing the value of the associated parameter.

The new value will correspond to the smallest activation level that meets equation 5.5

The same procedure is followed to increase the level when FMT is already activated.

If the state of our channel is close to the detection threshold, small fluctuations may

cause the FMT switch from one level of activation to the following. Frequent switch-

ing have a negative influence on the overall system performance and they increase the

amount of FMT signalling data required if FMT decision is not done locally. One way

of minimizing this problem is adding an hysteresis margin that will be apply when

56

Decision

the systems switch to a lower FMT level:

Systemmargin ≥ Detectionmargin +Hysteresis (8.6)

Accordingly we have to take the decision.In this way the decision is taken to counter-

act the fade.

8.2 Decision making algorithm

The flow chart for decision making is shown below. The inputs for decision making

is the predicted values of the channel that is the channel values are detected at time T,

and the channel values are predicted for a short time ahead. Based on this predicted

value the decision will be taken. The fade mitigation of our interest is adaptive coding

and modulation. The flow chart for decision has been presented in fig. 8.2.

In the simulation, the channel is modeled as a slow-varying Rayleigh channel with

Figure 8.2: Decision making flow chart

the negligible Doppler effect. The output of the channel is taken to be the attenuation

values in the simulation. Based on this attenuation values the SNR is calculated. The

57

Decision

detection algorithm presented in appendix will detect the SNR and estimate the SNR

a short time ahead. This SNR is given as input to the decision algorithm first the

algorithm checks the difference between the required SNR and the estimated SNR.

The difference is nothing but the margin that is available if margin is positive then we

do not require any FMT. If the margin is negative then we need to activate the FMT.

Now the algorithm will check for resources that is it will check whether resources are

available or not. If there are no resources available there will be outage that is the link

will fail. If resources are available then it will activate FMT.

Here the FMT of interest is adaptive coding and modulation hence it will activate it.

Now based on the amount of margin it will select the modulation scheme and coding

rate. If there are less impairments then we will go for higher order modulation scheme

and if there are more impairments we will go for lower order modulation scheme. In

this way decision is taken.

58

Decision

8.3 Results




ferent. Some of the curves are shown fig. 8.3 In the fig. 8.3 the x-axis is SNR and y

-axis represents the corresponding probability of errors. As we all know that as SNR

increases the corresponding BER will decreases. The different curves are for different

code rates for QAM, QPSK and BPSK and considered with AWGN channel.

−2 0 2 4 6 8 10 12 14 16 18

10−4

10−3

10−2

10−1

BER performance for 4−QAM(similar to QPSK), 16−QAM and 64−QAM

SNR (in dB)

BE

R

BPSK simulatedBPSK theoreticalQPSK simulatedQPSK theoretical16−QAM simulated16−QAM theoreticalBPSK 1/2 rate codeBPSK 1/3 rate codeQPSK 1/2 rate codeQPSK 1/3 rate code16−QAM 1/2 rate code16−QAM 1/3 rate code

Figure 8.3: SNR verses probability of error curves

8.3.2 Performance of the system for the collected data

The performance of the system for the collected data has been presented below. We

have simulated the system for the collected data that is the rain attenuation data has

been collected for several days and experiments have been performed on these data.

Basically the results have been presented for two days one is on 5th where we have

huge amount of rain and second is on sixth wherewe do not have any rain attenuation.

Change of modulation scheme and coding rate with respect to time

Based on the value of channel conditions the modulation scheme and coding rate

should be changed. If the channel condition is poor we will go for lower order mod-

ulation schemes and if the channel condition is good then we will go for higher order

59

Decision

modulation schemes.

0 200 400 600 800 1000 1200 1400 1600 18000

2

4

6

8

10

12

14

16

Time in sec

SN

R


Figure 8.4: change of modulation and coding with time(date 6th)

Actually the detection part has to detect the amount of fade present in the atmosphere

but we have using the rain data for simulation. The measurement of beacon signal

have been given that is the signal strength value is given based on which we need to

take the decision. Based on the values of signal strength we have selected the modula-

tion scheme and coding. In the figure. 8.4 we can observe that as the channel condition

is changing the modulation and coding rate are also changing. The bit error rate for

different modulation schemes with respect to time is presented in the figure. 8.7.

In figure. 8.5 we can observe that during 60 -80 sec there is a heavy rain because of

which the signal strength goes down which in turn increments the BER. By using

0 20 40 60 80 100 120 140 160 180−20

−15

−10

−5

0

5

10

15

20

Time in sec

SN

R


Figure 8.5: change of modulation and coding with time(date 5th)

60

Decision

ACM we can send the data but even then we will not have the enough margin to

overcome this outage. During that time we can go for some other FMTs like power

control in which we will increase the transmit power or site diversity in which link is

made available through some other terrestrial link. Joint FMTs will further improve

the system performance.

BER performance for the collected data

0 20 40 60 80 100 120 140 160 18010

−4

10−3

10−2

10−1

100

Time in sec

BE

R

M=2,C=1M=4,C=1M=16,C=1M=16, C=2BER, AMC

Figure 8.6: change of ber with time (with collected data(date 5th))

As the time varies the channel condition changes hence the bit error rate will also

changes. The change in BER at different point of times is shown in fig. 8.7. The thresh-

old BER that we have set was 10−1 and when we observe the simulated BER most of

the times it is below the target threshold if we use adaptive modulation and coding.

In the result we have shown the bit error rate for different modulation schemes. If we

use 16 QAM with 1/2 code rate we will achieve higher bit rate but we can observe

from the result that most of the times its bit error rate is above the threshold. Hence

if we use this constant rate transmission scheme then we will have outage most of the

times, which is not desirable. Same is the case with 16 QAM with rate 1/3. If we use

QPSK with code 1/2 and QPSK with no code then we can achieve the required bit

error rate but the data rate for these schemes will be less. So to achieve the higher data

rate and the required bit error rate we will go for adaptive coding and modulation,

that is based on the channel conditions we will change our modulation scheme and

coding rate. If the channel condition is poor then we will use lower order modulation

schemes and if the channel condition is good we will go for higher order modulation

61

Decision

0 200 400 600 800 1000 1200 1400 1600 180010

−4

10−3

10−2

10−1

100

Time in sec

BE

R

M=4,C=1M=2,C=1M=16,C=1M=16,C=2BER, AMCThreshold

Figure 8.7: change of ber with time (with collected data(date 6th))

schemes. the result is shown as BER, AMC in fig. 8.7. Hence we can conclude that by

using adaptive modulation and coding rate we can improve the data rate providing

the required bit error rate.

62

Decision

8.3.3 PER performance for different SNR values

The performance of the system with respect to packet error rate has been done here.

The packet size is taken as 1 Kb. With respect to the figure 8.8, it is seen that, when the

SNR is high, the per is less and the higher order modulation schemes and coding rates

are chosen. But, when the channel condition is worse, the lesser modulation schemes

and coding rates are applied so as to send as much less number of bits as possible.

0 5 10 15 20

10−1

100

PER performance for different ACM schemes

SNR (in dB)

PE

R

QPSK−1/2 simulatedQPSK−1/3 simulatedQPSK theoretical16−QAM simulated16−QAM−1/2 simulated16−QAM−1/3 simulated16−QAM theoretical64−QAM simulated64−QAM−1/2 simulated64−QAM−1/3 simulated64−QAM theoretical16−QAM theoretical16−QAM−1/2 simulated16−QAM−1/3 simulated64−QAM−1/2 simulated64−QAM−1/3 simulated

Figure 8.8: PER versus SNR curves

8.3.4 Performance of the FMT system with time


sented below. The estimated SNR almost tracks the calculated SNR most of the time.

Based on the estimated SNR, the decision of adaptive modulation and coding scheme

is taken. The corresponding diagram is shown in 8.9

63

Decision

Figure 8.9: Time versus SNR and change of M and C

64

Chapter 9

Delay compensation

Most communications satellites are located in the Geostationary Orbit (GSO) at an

altitude of approximately 35,786 km above the equator. At this height the satellites go

around the earth in a west to east direction at the same angular speed at the earth’s

rotation, so they appear to be almost fixed in the sky to an observer on the ground.

The time for one satellite orbit and the time for the earth to rotate is 1 sidereal day

or 23 h 56 m 4 seconds. Radio waves go at the speed of light which is 300,000 km

per second. If you are located on the equator and are communicating with a satellite

directly overhead then the total distance (up and down again) is nearly 72,000 km so

the time delay is 240 ms.

A satellite is visible from a little less than one third of the earth’s surface and if you

are located at the edge of this area the satellite appears to be just above the horizon.

The distance to the satellite is greater and for earth stations at the extreme edge of

the coverage area, the distance to the satellite is approx 41756 km. If you were to

communicate with another similarly located site, the total distance is nearly 84,000 km

so the end to end delay is almost 280 ms, which is a little over quarter of a second.

Extra delays occur due to the length of cable extensions at either end, and very much

so if a signals is routed by more than one satellite hop. Significant delay can also be

added in routers, switches and signal processing points along the route.

9.1 Delay calculation

The physical layer adaptation will consists of selecting suitable modulation and cod-

ing schemes which maximize channel efficiency for a certain BER, according to the

currently experienced channel state by each receiver. Inorder to do this we need to do

periodical channel estimations, which should be performed at the receiver end. this

channel state information has to be given to transmitter via signalling. Observing such

channel estimations, the transmitter shall select the most suitable modulation and cod-

65

Delay compensation

Figure 9.1: Delay calculation block diagram

ing scheme. The total time delay involved in the simulation is shown in the diagram.

The source will generate the information bits and these bits are given to transmitter

modem. The delay involved in generating the databits and processing in the modem

will be named as transmitter processing time(Tp). Now these modulated bits will be

transmitted to satellite the delay involved in this is named as Tupwhich will be greater

than 120ms. The satellite will relay these bits and retransmit it to the receiver the de-

lay involved in the down link is Tdown which will be greater than 120ms. The delay

involved in receiver processing, where the bits are demodulated, is named as receiver

processing time(Rp). Now from the received bits the channel state information is de-

tected the delay involved in detection is Tdet. The detected channel state information

has to be given to the transmitter through some feedback the delay involved here is

Tfeedback. Based on the detected information the decision function will take the deci-

sion the delay involved in this is Tdec. So the total time delay involved in this is given

by

Total Time delay = Tp+Tup+Tdown+Rp+Tdet+Tfeedback+Tpr+Tdec

Due to the inherent delay of the satellite system considered (transparent GEO), the

channel estimation information is available at the transmitter approximately 240 ms

66

Delay compensation

(for a GEO satellite case) after the moment that the estimation was done at the re-

ceiver, i.e. by the time that the channel estimation information is used at the Gateway

for ModCod selection, it is by definition outdated.

Such delays in the availability of channel state information might cause mismatches in

the selection of suitable ModCods if the channel has significantly changed during the

240 ms that it takes to signal the information to the transmitter. For that reason, de-

lay compensation strategies are required in order to minimize the ModCod selection

mismatches due to system delay.

9.2 Delay Compensation strategies

The strategies followed are based on the processing of the received channel state infor-

mation at the transmitter, by identifying what is the tendency of the channel estimates

compared with pervious measurements, i.e. identifying if the SNR tends to increase

or to decrease. According to this tendency estimation, a margin is applied to the latest

received measurement that follows the same tendency. In other words, the received

SNR estimation at the Gateway is not directly used for ModCod selection; instead the

received SNR is added or subtracted a certain margin if the channel tends to rise or

to fall, respectively. This emulates that the applied SNR for ModCod selection corre-

sponds to the current channel state at the time that the measurement is available.

In order to estimate if the channel tends to rise, to fall or to remain constant, each per-

formed channel estimation is compared with an old one. The difference between both

measurements is interpreted as shown in table. According to this interpretation of the

Difference Interpretation

[-1 dB, 1 dB] Channel is constantLess than -1 dB Channel is fallingGreater than 1 dB Channel is rising

channel tendency, the different strategies are proposed:

1. Asymmetric Back-off: with this strategy, if the channel is rising, the estimated

channel value is increased by 1.5 dB, whereas if the channel is falling, it is de-

creased by 4 dB. This strategy selects a considerably pessimistic ModCod in case

of falling channel.

2. Symmetric Back-off: with this strategy, the same margin, 1.5 dB, is applied in

case of falling and rising channel, whereas for rising channel this value is added

to the measurement and for falling channel the margin is subtracted.

3. Adaptive back-off: with this strategy the calculated difference between both

67

Delay compensation

compared measurements is directly applied as back-off. This allows for adap-

tation to the steepness of the channel tendency.

9.3 Delay Compensation flow chart

Delay compensation flow chart has been presented in the fig. 9.2. The input to the

algorithm is the current SNR and previous SNR. From these SNRs the margin will

be calculated. Based on the margin the decision will be taken and new SNR will be

calculated. Based on the new SNR the decision function will take the decision that is

the modulation and coding will be selected.

Figure 9.2: Delay Compensation flow chart

68

Delay compensation

Figure 9.3: Delay Compensation flow chart using adaptive back off

9.4 Results

9.4.1 SNR estimation with CRC and delay compensation

The estimation of SNR in the simulator has been done using CRC32. The estimation

of SNR will come under detection part , but in detecting the current SNR the detection

function will make use of previous modulation and coding rate. if the previous mod-

ulation and coding rate are accurate then the present estimation will be better. So here

basically we are using CRC to detect the errors and the detected SNR will be given to

decision part. The decision can be taken either applying back off or without applying

any back off. That is applying back off is compensating the delay and the correspond-

ing literature is given in the previous chapter. Based on the flow chart it will add some

margin to the detected SNR. The decision will be taken based on this new SNR. The

results has been presented for two cases one is with no back off and the other is with

adaptive back off. The curves are shown in fig 9.4. From the results we can conclude

that adaptive back off is giving more accurate estimation than no back off.

SNR estimation curves with symmetric back off ,asymmetric back off and adaptive

back off have been presented in 9.6. From the results we can observe that adaptive

69

Delay compensation

back off gives the better results when compared with the symmetric and asymmetric

curves.

Figure 9.4: Comparison of SNR and data rate curves with and without back off for CRC

70

Delay compensation

Figure 9.5: Comparison of SNR and data rate curves with and without back off for CRC

Figure 9.6: Comparison of SNR curves with and without back off for CRC

71

Delay compensation

9.4.2 SNR estimation with Continuous pilot and delay compensa-

tion

Figure 9.7: Comparison of SNR and data rate curves with and without back off for continuouspilot

The channel estimation can be done with embedded pilot. That is along with data

some pilot bits are transmitted to estimate the channel values. These pilot bits have to

be transmitted at regular intervals. If the pilot bits are transmitted continuously that

is all at a time then it is said to be continuous pilot. In the fig 9.7 continuous pilot

channel estimation is shown. The results show that adaptive back-off is giving more

accurate estimation than no back-off.

72

Delay compensation

20 40 60 80 100 120 140−2

0

2

4

6

8

10

12

time unit in sec

PE

R d

ecis

ion

snr and data rate curves for continuous pilot with and without backoff

SNR estimated with no back offSNR calculatedSNR estimated with adaptive back offDATA RATE OPTIMUMDATA RATE WITH nbfDATA RATE WITH ADAPTIVE BACK OFF

Figure 9.8: Comparison of SNR and data rate curves with and without back off for continuouspilot

Figure 9.9: Comparison of SNR curves with and without back off for Continuous pilot

73

Delay compensation

9.4.3 SNR estimation with Distributed pilot and delay compensa-

tion

In the distributed pilot the pilot bits are distributed among the data bits. The channel

values can be estimated with this distributed bits. the curves are shown in fig 9.10.

SNR estimation curves with symmetric back off, asymmetric back off and adaptive

back off have been presented in 9.12. From the results we can observe that adaptive

back off gives the better results when compared with the symmetric and asymmetric

curves.

Figure 9.10: Comparison of SNR and data rate curves with andwithout back off for distributedpilot

74

Delay compensation

20 40 60 80 100 120−2

0

2

4

6

8

10

12

time unit in sec

PE

R d

ecis

ion

snr curves for distributed pilot with and without back off

SNR calculatedDATA RATE OPTIMUMSNR estimated with NBFSNR estimated with adaptive back offDATA RATE WITH NBFDATA RATE with adaptive back off

Figure 9.11: Comparison of SNR and data rate curves with andwithout back off for distributedpilot

Figure 9.12: Comparison of SNR curves with and without back off for Distributed pilot

75

Chapter 10

Results

The adaptive delay has being discussed in the previous chapter.SNR Moving average

is a technique used to analyze time series data, in which a weighted average is deter-

mined for a given data point based on its value and the past values. The results shown

can mainly be divide into four categories:-

10.1 Without SNR moving average

10.1.1 Without SNR moving average and adaptive back off

Comparison of the BLER and Throughput curves for without SNR moving average

and adaptive back off.

Figure 10.1: Comparison of BLER and SNR curvesWithout SNRmoving average and adaptiveback off

76

Results

10.1.2 Without SNR moving average and no adaptive back off

Comparison of the BLER and Throughput curves for without SNR moving average

and no adaptive back off.

Figure 10.2: Comparison of BLER and SNR curvesWithout SNRmoving average and no adap-tive back off

Figure 10.3: Comparison of Throughput and SNR curves Without SNR moving average andno adaptive back off

77

Results

10.2 With SNRmoving average

10.2.1 With SNR moving average and adaptive back off

Comparison of the BLER and Throughput curves for with SNR moving average and

adaptive back off.

Figure 10.4: Comparison of BLER and SNR curves With SNR moving average and adaptiveback off

Figure 10.5: Comparison of Throughput and SNR curvesWith SNRmoving average and adap-tive back off

78

Results

10.2.2 With SNR moving average and no adaptive back off

Comparison of the BLER and Throughput curves for with SNR moving average and

no adaptive back off.

Figure 10.6: Comparison of BLER and SNR curvesWith SNRmoving average and no adaptiveback off

Figure 10.7: Comparison of Throughput and SNR curves With SNR moving average and noadaptive back off

79

Chapter 11

Updated Results

11.1 Graphs

The results shown here are for SNRmoving average along with three different back-off

schemes. The results are presented below:-

2 4 6 8 10 12 14 16 18 20

1

2

3

4

5

6x 10

6

SNR in dB

Th

rou

gh

pu

t in

bits

/se

c

dist pilot MA ADBF offset 3 dbcont pilot MA ADBF offset 3 dbcrc MA ADBF offset 3 dbcrc MA SBF offset 3 dbDist pilot MA SBF offset 3 dbCont pilot MA SBF offset 3 dbcrc MA ADBF offset 2 dB

Figure 11.1: Throughput and SNR curves of different Back-Off schemes

80

Updated Results

CRC dist cont0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

me

an

BL

ER

fo

r d

iffe

ren

t sc

he

me

s

ADBF 2dB offsetSBF 2dB offsetADBF 3dB offsetSBF 3dB offsetMA NBF 2 dB offset

Figure 11.2: Error performance of different Back-Off schemese

81

Updated Results

CRC Dist pilot Cont pilot0

2

4

6

8

10

12

14

16x 10

5

me

an

of

thro

ug

hp

ut(

bp

s) f

or

diff

ere

nt

sch

em

es

SBF 3dB offset

ADBF 3dB offset

ADBF 2dB offset

SBF 2dB offset

Figure 11.3: Throughput performance of different Back-Off schemes

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

BLER

CD

F o

f B

LE

R

crc MA ADBF offset 3 db

cont pilot MA ADBF offset 3 db

dist pilot MA ADBF offset 3 db

cont pilot MA SBF offset 3 db

dist pilot MA SBF offset 3 db

crc MA SBF offset 3 db

crc MA ADBF offset 2 dB

Figure 11.4: CDF of BLER for different Back-Off schemes

82

Updated Results

0 1 2 3 4 5 6

x 106

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Throughput

CD

F o

f T

hro

ug

hp

ut

CRC MA ADBF offset 3 dbDist pilot MA ADBF offset 3 dbCont pilot MA ADBF offset 3 dbCont Pilot MA SBF offset 3 dbDist Pilot MA SBF offset 3 dbcrc MA SBF offset 3 dbcrc MA ADBF offset 2 dB

Figure 11.5: CDF of Throughput perforomance of different Back-Off schemes

83

Updated Results

5 10 15 20 25 30 35 40 45 50 55

0.7

0.75

0.8

0.85

0.9

0.95

Time in secs

Cro

ss−

corr

ea

ltio

n v

alu

e o

f S

NR

(ca

l an

d e

st)

CRC MA ADBF offset 3 db

Cont pilot MA ADBF offset 3 db

Dist pilot MA ADBF offset 3 db

Cont pilot MA SBF offset 3 db

Dist pilot MA SBF offset 3 db

crc MA SBF offset 3 db

crc offset 2 dB MA ADBF

Figure 11.6: Cross-Correlation between the calculated and estimated SNR

84

Chapter 12

Practical Implementation with Modem

(SRM6100)

12.1 Introduction of Modem-SRM6100

The SRM6100 transceiver modems are high performance wireless radio modems de-

signed for heavy-duty industrial data communications in the 2.4- 2.4835 GHz license-

free band. The SRM6100 can be operated in a number of different modes to satisfy a

broad range of communications requirements. It can be configured for point-to-point

or multipoint operation with a unlimited number of remote sites on a single master

depending on data throughput requirements. The SRM6100 will operate in virtually

any environment where RS232 data communications are required. The transceiver

RS232 interface is a standard DB9F connector that is configured for Data Communica-

tions Equipment (DCE) operation. The SRM6100 will connect with a straight through

RS232 cable to a device configured for Data Terminal Equipment (DTE) operation.

12.2 Specifications of Modem-SRM6100

1. This modem is a transceiver modem with a high performance wireless radio de-

signed for heavy-use industrial data communications in the range of 2.4-2.4835

GHz .

2. Advanced frequency hopping and error detection technology to provide data

integrity

3. The modem has data-rates of 1200-234.4kbps.

4. The operating range of the modem is in the range of 24 km in optimal conditions

of line-of-sight[12].

85

Practical Implementation with Modem (SRM6100)

12.3 Experiments done with the modem SRM6100

12.3.1 Loop Back Bench Test

In this test,the data was transmitted from one transmitter PC to the Receiver by the

hyper terminal. The test was done by following the instructions from the manual. In

the receiver modem the receive and transmit terminals were shorted so that what ever

data was being received at the receiver was transmitted back to the transmitter hence

it is called lopp back bench test. The details of the test are as follows:

1. Attach the bench test antenna included with the radio modem

2. Locate the SRM6100 labeledMASTER. Using a standard RS232 cable, connect the

radiomodem to a communication port on a computer that has a communications

utility such as HyperTerminal, ProComm Plus or Terminal for Win3.x. Set the

data rate (BPS) of the terminal program to match the port rate of the SRM6100.

Plug the power supply into an AC outlet of the correct voltage and connect the

power supply to the SRM6100. The red LED marked P (power) on the radio

modem front panel should turn on.

3. Locate the SRM6100(s) labeled REMOTE. Connect the power supply to the SR-

M6100. The red LED marked P (power) on the REMOTE radio modem should

turn on. If it is a point-to-point system the amber LED C (carrier detect) should

turn on for both the Remote and master. If it is a point to multipoint system the

C LED will turn on for the Remote only. Attach a Loop Back test jumper on the

RS232 data DB9F connector of the SRM6100 remote. The jumper shorts pins 2

and 3 of the data connector.

4. Using the terminal that is connected to the MASTER SRM6100, hold down a key,

A for example. The letter A should begin to scroll across the terminal screen.

This indicates that the data (the letter A in this case) is being transmitted from

the terminal through the MASTER SRM6100, through the REPEATER (if appli-

cable), on to the REMOTE SRM6100, through the Loop Back test jumper, back

through the REPEATER (if applicable) to the MASTER, and then to the termi-

nal. This establishes that the SRM6100s are functioning in full duplex mode and

are operating properly. If something appears scrolling across the terminal screen

other than the correct character for the key being pressed, it indicates that the

terminals settings and data rate may not be set to match that of the SRM6100.

5. While continuing to press the letter A the yellow LED marked I (Input) and the

green LED marked O (Output) should both be flashing rapidly on the Master

radio modem and the Remote with the jumper attached. Remove the jumper

86


from the REMOTE radio modem. The letter display scrolling across the screen

should stop, and the O LED will stop flashing at the MASTER The I LED will

flash each time the key is pressed; indicating that the radio modem is receiving

a data input signal on the RS232 port. The O LED on the REMOTE will flash

each time a key is pressed; indicating that the radio modem is outputting a data

input signal on the RS232 port. The I LED on the REMOTE will remain off with

no data loop back. Replace the Loop Back test jumper in the REMOTE radio

modem. Hold down the key again, and the letter should once again scroll across

the computer screen. If there is a REPEATER in the system its C (carrier detect)

LED will flash rapidly when data is being passed. The REPEATER I and O LEDs

remain off during normal operation.

The test has been performed in HyperTerminal as well as Matlab.

12.3.2 Configuration setting

The SRM6100 allows to set several parameters to suit particular application. All ad-

justments are done through the SRM6100 setup program. To access the configuration

menu, connect the radio modem to terminal programwith port settings of 19.2 Kbaud,

8 data bits, no parity and one stop bit. With the modem connected to the PC running

the terminal program, press the Configure button located behind the pinhole next to

the DB9 connector on the front of the modem. once we press the configure button we

will get the main menu. The main menu provides the radio modems unique call book

number and the set of choices for editing the operational parameters and viewing the

performance data.

Data transmission with Configuration setting

The modem’s were interfaced with Matlab. the configuration settings of the modems

were changed through the Matlab program. The configuration settings include the

transmit power and data rate. First the data has been transmitted then configuration

settings have been changed that means the next data will be transmitted with this new

configuration settings. At the end the radio statistics has been displayed.

12.4 Limitations of modem SRM6100

1. Themodem should always be triggeredmanually to show the configuration win-

dow

2. There are not enough options in SRM6100 to change the modulation scheme and

coding rate

87


3. The actual data rate is not known to the user and thus there are not many data

rate options.

88

Chapter 13

Plan of the experiment had to be done

at SAC,Ahmedabad

13.1 Objective of the experiment

To measure the unlock-to-lock delay performance of the high-speed satellite modem

13.2 Scope

1. The modem used in the experiment is the CDM-700 Satellite Modem.

2. We will test the unlock-to-lock delay, mainly, by changing the Modulation and

Coding (M,C) scheme. This will be done by giving the corresponding command

to the two modems. We will also try to study the effect of changing the attenua-

tion value in the attenuator on the delay value.

3. The switching delay occurring while sending data to the Tx-modem and getting

data from the Rx-modem will be considered negligible for the purpose of our

experiment. Basically, we will try to find an upper-bound to the time the system

takes to change the M,C scheme and get ready to send and receive data.

13.3 Experimental Set-up

The following connections are necessary:

1. The data-bits to be transmitted are sent from the computer to the transmitter-

modem

2. The Rx-modem sends the total received bits to the computer for BER calculation

89

Plan of the experiment had to be done at SAC,Ahmedabad

Figure 13.1: plan of the experiment

3. Connection for the attenuator with the computer

4. Connection for transmitting the M,C decision

5. Connection to the receiver-modem regarding the M,C that has been selected at

the transmitter-modem.

6. IF/RF Connection for transmission of the modulated bits from the transmitter to

the attenuator.

7. IF/RF Connection for transmission of the modulated bits from the attenuator to

the receiver modem.

13.4 Methodology

1. The computer is the information source in our experiment. The data is gener-

ated as a random sequence of bits and is sent to the Transmitter-modem by data

interface (1).

2. Initially, the experiment is started with a pre-defined modulation and coding

rate. This information is given to the transmitter modem (4) and receiver modem

(5) by the computer. Afterwards, the modulation and coding rates are selected

based on the BER calculation.

90


3. The attenuator models the varying channel in the entire experiment. In the above

block diagram, the attenuator is programmable by the system. The transmitted

bits are attenuated by the attenuator block (3). In case, a programmable attenu-

ator is not available, a manual attenuator will also do. But in that case, we will

have to change the attenuation values manually by hand. If an attenuator is not

available at all, we just have to send the transmitted signal to the receiver. Nat-

urally we would not be able to study the effect of channel attenuation on the

unlock-to-lock delay in this case.

4. The received data-bits are transferred from the receiver-modem to the computer

(2).

5. Based on the BER results, the M,C are changed. This change is intimated to

the two modems which subsequently change their M,C schemes. When the

”change” message is sent to the two modems, two counters are started - one

for the TX-modem and the other for the Rx-modem. The two modems are then

queried continuously to find whether they have been able to lock or not. When

any of the two is locked, the corresponding counter is stopped and the time-

delay is noted down.

6. Steps 1-5 are repeated for a sufficient number of times for finding the delay per-

formance of the modem.

13.5 Resources required

1. Computer

2. Transmitter and Receiver Modems - both CDM-700 Satellite Modems

3. Programmable Attenuator interfaced to the computer (optional)

4. Data interface (for the two modems)

13.6 Expected Results

1. Estimation of the average and maximum values of delay between sending the

M,C scheme information to the modems and getting the modems ready to send

and receive data after M,C switching.

2. Finding whether changing the channel attenuation has any effect on the unlock-

to-lock delay. If it has, what is the effect like?

91


3. Implementation of the FMT loop operating in real-time. This will be possible

only if the attenuator is programmable using the computer.

92

Chapter 14

Conclusion

14.1 Conclusion

In order to implement the fade mitigation techniques in satellite communication at

higher frequency bands, we need to know the various kinds of impairments and their

impact on the satellite links firstly. The channel fade should be calculated and accord-

ingly decision should be taken. Once the channel fade is known,decision for activation

of fade mitigation technique is to be taken. Then we can activate the FMT’s as and

when needed to have a good performance of the link.

From the results it is observed,that, adaptive modulation and coding perfroms bet-

ter than fixed rate transmission schemes.The ACM works well but it may not remove

the outage everytime. Some times we need to use the other FMTs when there is not

enough resources for ACM. Combination of different FMT’s will further improve the

system performance. The delay compensation strategies have been applied to the sys-

tem. The methods include symmetric back off, asymmetric back off and adaptive back

off. Amongst all adaptive back off is has better results for both CRC and Embedded

pilots The simulation results show that the Adaptive back-off performs better than the

other back-off schemes as it is a type of scheme that adapts to the slope of the chan-

nel fluctuations. The CRC scheme has a better performance given the SNR offset to

be 2 dB. For the ADBF case, CRC satisfies the BLER almost all the time(only 1.05% it

doesnt satisfy). But for the pilots,the BLER does not satisfy for 23%-24% of the time,to

be precise. So,in a nut shell,the CRC performs better than the embedded pilots given

the SNR offset of 2dB.

14.2 Future scope

The system is simulated for slow varying rayleigh channel. Instead of that, a rain fade

model has to be developed. A prediction algorithm has to be developed and to be

93

Conclusion

included into the system to get better results. The whole system is to be implemented

on hardware to verify the system performance. The nonlinearities of high power am-

plifier should also be studied in detail.

94

Appendix A

List of Abbreviations

GEO Geostationary Earth Orbit

FMT Fade Mitigation Techniques

SNR Signal to Noise Ratio

TEC Total Electron Content

ULPC Up link Power Control

DLPC Down link Power Control

EEPC End-to-End Power Control

OBBS On-Board Beam Shaping

EIRP Effective Isotropically Radiated Power

FEC Forward Error Correction

SD Site Diversity

FD Frequency Diversity

FCM Fade Counter Measure

AMC Adaptive Modulation and Coding

BER Bit Error rate

PER Packet Error rate

VSAT very small aperture terminals

DRR Data Rate Reduction

AM Adaptive Modulation

AC Adaptive Coding

95

List of Abbreviations

ARQ Automatic Repeat Request

QAM Quadrature Amplitude Modulation

AWGN Additive White Gaussian Noise

BPSK Binary Phase Shift Keying

QPSK Quadrature Phase Shift Keying

CRC Cyclic Redundancy Check

AGC Automatic Gain Control

LNA Low Noise Amplifier

96

Bibliography

[1] Laurent CASTANET, Ana BOLEA-ALAMANAC, Michel BOUSQUET; ”Inter-

ference and fade mitigation techniques for Ka and Q/V band satellite commu-

nication systems”

[2] Laurent CASTANET, Ana BOLEA-ALAMANAC, Michel BOUSQUET, Max

VAN DE KAMP; ”Implementation of short-term prediction models in fade mit-

igation techniques control loops” Joint COST 272 / 280 Workshop,26-28 May

2003, ESA/ESTEC, Noordwijk, The Netherlands

[3] L. Castanet,J. Lemorton,M. Bousquet;A ”Modular simulation tool of interference

and fade mitigation techniques applied to millimeter wave satcom systems”

COST Action 280, 1st International Workshop

[4] Timothy Pratt, Charles Bastian, Jeremy Allnutt; ”Satellite communications” sec-

ond edition

[5] Anbazhagan Aroumont, Ana-Bolea Alamanac, Laurent Castanet, and Michel

Bousquet; ”SNR estimation algorithms for Channel Quality detection in

Ka/EHF band Satellite systems employing Fade Mitigation Techniques”

[6] Dr. Louis J. Ippolito ; ”Propagation Effects Handbook for Satellite Systems De-

sign”

[7] Prof.S.S.Das; ”Techniques to enhance the spectral efficiency of OFDM wireless

systems” Alabog university, Center for TeleinFrastructer,7 september 2007

[8] B.C.Gremont; ”Event detection, control and performance modeling”

97

BIBLIOGRAPHY

[9] M.M.J.L van de Kamp:”Rain Attenuation As a Markov Process: How To Make

an Event”,ONERA-DEMR, Toulouse, France.

[10] M.M.J.L van de Kamp: ”Rain Attenuation as a Markov Process:The Meaning

of Two Samples”, COST Action 280, ”Propagation Impairment Mitigation for

Millimetre Wave Radio Systems”

[11] Maseng, T. and P. M. Bakken: ”A Stochastic Dynamic Model of Rain Attenua-

tion”,IEEE Trans. Commun., COM-29(5), 660-669, 1981.

[12] User Guide,SRM6100: ”Frequency Hopping Spread Spectrum Radio Modem”

[13] Cristina Parraga Niebla, Christian Kissling: ”Delay Compensation Strategies for

an Efficient Radio Resource Management in DVB-S2 Systems”.

[14] Michele Luglio: ”Fade Prediction and Control systems” IJniversid di Roma Tor

Vergata - Dipartimento di Ingegneria Elettronica Via della Ri.cerca Scientifica

00133 - Roma - Italy

[15] Dieter Kreuer+, Ansgar Radermacher+, Sven Martin*, Thomas Frach+: ”A Pro-

totype Adaptive Fade Countermeasure System for TDMA Operation at Ka

Band” +Institute of Computer Science IV, *Institute of Computer Science VI,

Aachen University of Technology D-52056 Aachen, Germany

[16] Jun Sun*, Jay Gao+, Shervin Shambayati+, and Eytan Modiano*: ”Ka-Band Link

Optimization with Rate Adaptation” *-Laboratory of Information and Decision

Systems , Massachusetts Institute of Technology Cambridge, MA 02139 +Jet

Propulsion Laboratory California Institute of Technology 4800 Oak Grove Drive,

Pasadena, CA 91109

[17] Andrei Malygin, Miodrag Filip and Enric Vilar: ”Neural network implementa-

tion of a fade countermeasure controller for a VSAT link” INTERNATIONAL

JOURNAL OF SATELLITE COMMUNICATIONS Int. J. Satell. Commun. 2002;

98

BIBLIOGRAPHY

[18] Seong Taek Chung, Student Member, IEEE, and Andrea J. Goldsmith, Senior

Member, IEEE: ”Degrees of Freedom in Adaptive Modulation: A Unified View”

IEEE TRANSACTIONS ON COMMUNICATIONS, VOL. 49, NO. 9, SEPTEM-

BER 2001

[19] Prof. Kalyan BANDYOPADHYAY IIT Kharagpur, INDIA October 2, 2008

IWSSC: ”GSAT-4, a step towards Indian Advance Communications Satellite”

[20] T K P Chung, A P Gallois: ”Fade countermeasure intelligent system”

[21] K.S. Dasgupta, A. Charania, A. Shukla, R. Acharya, K. Bandyopadhyay, L. Cas-

tanet, J. Lemorton, G. Carrie, F. Lacoste, J.P. Taisant, F. Carvalho: ”A new prop-

agation campaign in tropical areas: the Ka-band Propagation experiment over

India with the GSAT-4 satellite”

[22] ATHANASIOS D. PANAGOPOULOS, PANTELIS-DANIEL M. ARAPOGLOU,

AND PANAYOTIS G. COTTIS: ”Satellite Communications at Ku, Ka, And V

Bands:Propagation Impairments and mitigation techniques”

[23] Neno Celandrino, Erina Ferro and Francesco Potorti: ”Quality Estimation of

PSK Modulated Signals”

[24] :Rain Attenuation As a Markov Process: How To Make an Event,ONERA-

DEMR, Toulouse, France.

[25] Rain Attenuation as a Markov Process:The Meaning of Two Samples, COST Ac-

tion 280, Propagation ImpairmentMitigation forMillimetreWave Radio Systems

[26] T. and P. M. Bakken: A Stochastic Dynamic Model of Rain Attenuation, IEEE

Trans. Commun., COM-29(5), 660-669, 1981.

99

Documents

Studies on Fade Mitigation Control for Microwave …gssst.iitkgp.ac.in/GSProjects/SFM/reports/SFM_main...Date:November 10, 2010 Acknowledgement I would like to thank my project supervisor