Upload
others
View
14
Download
0
Embed Size (px)
Citation preview
Distributed Detection and Estimation in Wireless Sensor Networks:Resource Allocation, Fusion Rules, and Network Security
Edmond Nurellari
The University of Leeds, UKSchool of Electronic and Electrical Engineering
In accordance with the requirements for the degree of Doctor of Philosophy
June 6, 2017
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 1 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 2 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 2 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 2 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 2 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 2 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 2 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 2 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 3 / 42
1. Introduction
Motivation
WSNs spatially deployed over a field can be designed to collect information and monitormany phenomena of interest.
Important role in several daily application scenarios such as health-care monitoring, homeapplications, smart farming, environment monitoring, and military.
1.2. Design Challenges in WSNs
1
FUSIONCENTERAttacker
. . .
SN1
SN2
SN3
SN4
SN5
SN6
SN7
SN8
SN9
SN10
Nature
T1
T2
T3
T4
T6
T5
SN1
SN3
SN10
SN8
←→−→
active sensor node manipulated by the attacker
active sensor node, working accordingly
non-active sensor node, manipulated by the attacker
non-active sensor node, controlled by the fusion center
partial inter-sensor node connectivity
target to SN link
SN1
SN3
SN10
SN8
←→−→
active sensor node, manipulated by the attacker
active sensor node, working accordingly
non-active sensor node, manipulated by the attacker
non-active sensor node, controlled by the fusion center
partial inter-sensor node connectivity
target to SN link
Figure 1.1: Schematic for a distributed communication architecture among periph-
eral SNs. Each SN generates a test statistic (Ti) by observing the target (thick lines).
The SNs have partial connectivity (thin lines) among themselves (i.e., not a complete
graph), but only over an energy-constrained/bandwidth-constrained network.
3
Figure 1: (left) A WSN architecture. (right) Smart city infrastructure.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 4 / 42
1. Introduction
Motivation
WSNs spatially deployed over a field can be designed to collect information and monitormany phenomena of interest.
Important role in several daily application scenarios such as health-care monitoring, homeapplications, smart farming, environment monitoring, and military.
1.2. Design Challenges in WSNs
1
FUSIONCENTERAttacker
. . .
SN1
SN2
SN3
SN4
SN5
SN6
SN7
SN8
SN9
SN10
Nature
T1
T2
T3
T4
T6
T5
SN1
SN3
SN10
SN8
←→−→
active sensor node manipulated by the attacker
active sensor node, working accordingly
non-active sensor node, manipulated by the attacker
non-active sensor node, controlled by the fusion center
partial inter-sensor node connectivity
target to SN link
SN1
SN3
SN10
SN8
←→−→
active sensor node, manipulated by the attacker
active sensor node, working accordingly
non-active sensor node, manipulated by the attacker
non-active sensor node, controlled by the fusion center
partial inter-sensor node connectivity
target to SN link
Figure 1.1: Schematic for a distributed communication architecture among periph-
eral SNs. Each SN generates a test statistic (Ti) by observing the target (thick lines).
The SNs have partial connectivity (thin lines) among themselves (i.e., not a complete
graph), but only over an energy-constrained/bandwidth-constrained network.
3
Figure 1: (left) A WSN architecture. (right) Smart city infrastructure.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 4 / 42
1. Introduction
Motivation
WSNs spatially deployed over a field can be designed to collect information and monitormany phenomena of interest.
Important role in several daily application scenarios such as health-care monitoring, homeapplications, smart farming, environment monitoring, and military.
1.2. Design Challenges in WSNs
1
FUSIONCENTERAttacker
. . .
SN1
SN2
SN3
SN4
SN5
SN6
SN7
SN8
SN9
SN10
Nature
T1
T2
T3
T4
T6
T5
SN1
SN3
SN10
SN8
←→−→
active sensor node manipulated by the attacker
active sensor node, working accordingly
non-active sensor node, manipulated by the attacker
non-active sensor node, controlled by the fusion center
partial inter-sensor node connectivity
target to SN link
SN1
SN3
SN10
SN8
←→−→
active sensor node, manipulated by the attacker
active sensor node, working accordingly
non-active sensor node, manipulated by the attacker
non-active sensor node, controlled by the fusion center
partial inter-sensor node connectivity
target to SN link
Figure 1.1: Schematic for a distributed communication architecture among periph-
eral SNs. Each SN generates a test statistic (Ti) by observing the target (thick lines).
The SNs have partial connectivity (thin lines) among themselves (i.e., not a complete
graph), but only over an energy-constrained/bandwidth-constrained network.
3
Figure 1: (left) A WSN architecture. (right) Smart city infrastructure.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 4 / 42
1. Introduction
Design Challenges in WSNs
Low Power Hardware: Clearly, the biggest design constraint in WSNs still remains thepower consumption. Even-though the SNs are being designed using low-power microcontrollers, their power dissipation is still orders of magnitude too high.
Resource Constraints: Battery operated devices with limited on-board energy, both thesystem lifetime and communication bandwidth (BW) are restricted. Both the signalprocessing and communication should be carefully designed to consume minimal energy inorder to extend the lifetime and improve the overall reliability of the WSN.
Network Security:Usually unattended (geographically dispersed) and this makes themvulnerable to attacks. The overall detection and estimation strongly depends on thereliability of these SNs.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 5 / 42
1. Introduction
Design Challenges in WSNs
Low Power Hardware: Clearly, the biggest design constraint in WSNs still remains thepower consumption. Even-though the SNs are being designed using low-power microcontrollers, their power dissipation is still orders of magnitude too high.
Resource Constraints: Battery operated devices with limited on-board energy, both thesystem lifetime and communication bandwidth (BW) are restricted. Both the signalprocessing and communication should be carefully designed to consume minimal energy inorder to extend the lifetime and improve the overall reliability of the WSN.
Network Security:Usually unattended (geographically dispersed) and this makes themvulnerable to attacks. The overall detection and estimation strongly depends on thereliability of these SNs.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 5 / 42
1. Introduction
Design Challenges in WSNs
Low Power Hardware: Clearly, the biggest design constraint in WSNs still remains thepower consumption. Even-though the SNs are being designed using low-power microcontrollers, their power dissipation is still orders of magnitude too high.
Resource Constraints: Battery operated devices with limited on-board energy, both thesystem lifetime and communication bandwidth (BW) are restricted. Both the signalprocessing and communication should be carefully designed to consume minimal energy inorder to extend the lifetime and improve the overall reliability of the WSN.
Network Security:Usually unattended (geographically dispersed) and this makes themvulnerable to attacks. The overall detection and estimation strongly depends on thereliability of these SNs.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 5 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Contribution-Publications List
1 E. Nurellari, D. McLernon, and M. Ghogho “A Secure Optimum Distributed Detection Scheme in Under-AttackWireless Sensor Networks”, in IEEE Trans. on Signal and Information Processing over Networks (TSIPN), April2017.
2 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Binary Event Detection Under Data-Falsification andEnergy-Bandwidth Limitation”, in IEEE Sensors Journal, vol. 16, no. 16, pp. 6298-6309, Aug. 15, 2016.
3 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed Two-Step Quantized Fusion Rules via ConsensusAlgorithm for Distributed Detection in Wireless Sensor Networks,” in IEEE Transactions on Signal and InformationProcessing over Networks (TSIPN), vol. 2, no. 3, pp. 321-335, Sept. 2016.
4 S. Aldalahmeh, M. Ghogho, D. McLernon, and E. Nurellari, “Optimal fusion rule for distributed detection inclustered wireless sensor networks”, EURASIP Journal on Advances in Signal Process., 2016:5, Jan. 2016.
5 E. Nurellari, D. McLernon, and M. Ghogho, “Distributed detection in practical wireless sensor networks via a twostep consensus algorithm,” in Proc. IET Int. conf. on Intelligent Signal Process. (ISP), London, United Kingdom,1-2 Dec. 2015.
6 E. Nurellari, D. McLernon, M. Ghogho and S. A. R. Zaidi, “Distributed Optimal Quantization and PowerAllocation for Sensor Detection Via Consensus,” Proc. IEEE VTC Spring, Glasgow, U.K., 11-14 May 2015.
7 E. Nurellari, S. Aldalahmeh, M. Ghogho, and D. McLernon, “Quantized Fusion Rules for Energy-BasedDistributed Detection in Wireless Sensor Networks,” Proc. IEEE SSPD, Edinburgh, Scotland, 8-9 Sep. 2014.
8 E. Nurellari, D. McLernon, M. Ghogho and S. Aldalahmeh, “Optimal quantization and power allocation forenergy-based distributed sensor detection,” Proc. IEEE EUSIPCO, Lisbon, Portugal, 1-5 Sept. 2014.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 6 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 7 / 42
2. Optimal Quantization and Power Allocation
System Architecture
SN3
SN2
SN5
SN1
SN4
Target
Fusion Center
T qf =
M∑i=1
αi [Ti ]QT3
T2
T5
T4
T1
[T3]Q
[T2]Q
[T5]Q[T4]Q
[T1]Q
Figure 2: Communication architecture between peripheral SNs and the FC. Each SN generates a test statistic byobserving the target and can communicate with the FC only over an energy-constrained/bandwidth-constrained link.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 8 / 42
2. Simulation Results 1/2
1 2 3 4 5 6 7 8 9 100
0.5
1
1.5
sensor i
h2 i
1 2 3 4 5 6 7 8 9 100
2
4
sensor i
p i
1 2 3 4 5 6 7 8 9 100
2
4
sensor i
L i
Equal weighting in (3.3.4)
Optimum weighting in (3.3.4)
Figure 3: Equal weight (αi = 1√M,∀i) and optimal weight combining (α = αopt) transmit power and
channel quantization bits allocation for Pfa = 0.1, Pt = 10, U = 0.1, and M = 10.Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 9 / 42
2. Simulation Results 2/2
0 0.2 0.4 0.6 0.8 10
0.2
0.4
0.6
0.8
1
Probability of false alarm, Pfa
Prob
abilit
yof
dete
ctio
n,P d
Optimal weight, N=100 samplesOptimal weight, N=300 samplesEqual weight, N=100 samplesEqual weight, N=300 samples
Figure 4: Receiver operating characteristic with Pt = 10, U = 0.1 and M = 10 for two differentweighting schemes.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 10 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 11 / 42
3. Simulation Results 1/3
-14 -12 -10 -8 -6 -40
0.2
0.4
0.6
0.8
1
Prob
abilit
yof
dete
ctio
n.P d
a
(dB)
Opt LRT-based
LRT-based in (4.4.8)
Opt lin comb in (4.4.9)
Eq LRT-based
Linear combi in (4.3.9)
Eq lin combining
Figure 5: Probability of detection (Pd ) versus the signal to noise ratio (ξa) for M = 20, N = 10,Pt = 10, Pfa = 0.1 and B = 0.5.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 12 / 42
3. Simulation Results 2/3
0 20 40 60 80 100 1200
0.2
0.4
0.6
0.8
1
number of samples, N
Prob
abilit
yof
dete
ctio
n,P d
Opt fusion rule, Pt=102
Opt linear combining, Pt=102
Opt fusion rule, Pt=10-1
Opt linear combining, Pt=10-1
Figure 6: Probability of detection (Pd ) versus the number of samples (N) for M = 10 sensors,Pfa = 0.1, ξa = −8.5 dB and B = 1.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 13 / 42
3. Simulation Results 3/3
20 40 60 80 100
0.4
0.5
0.6
0.7
0.8
0.9
1
number of sensors, M
Prob
abilit
yof
dete
ctio
n,P d
Optimum fusion rule LRT-based
LRT-based with weights in (4.4.8)
Optimum linear combining in (4.4.9)
Equal weight LRT-based
Linear combining with weights in (4.3.9)
Equal weight linear combining
Figure 7: Probability of detection (Pd ) versus number of sensors (M) for N = 10, Pt = 10, Pfa = 0.1,ξa = −8.5 dB and B = 0.5.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 14 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 15 / 42
4. Distributed Two-Step Quantized Fusion Rules
Communication Architecture
SN1
SN2
SN3
SN5
SN4
SN6
Target
T1
T2
T3
T4
T6
T5
Figure 8: A distributed communication architecture among peripheral SNs. The SNs have partial connectivity (thinlines) among themselves (i.e., not a complete graph).
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 16 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
Here we propose a scheme, where SN i encodes the data (using a simple uniformquantizer with qi bits) prior to information exchange.
1 We also propose to establish a link between any two SNs i and j based on the (known) SNRat node j , i.e.
if SNRij < Υ, eij = eji = 0
if SNRij ≥ Υ, eij = eji = 1.
}
2 Υ is a SNR threshold parameter and SNRij defined as:
SNRij =pt
ijh2ij
ζ0dγij
.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 17 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
Here we propose a scheme, where SN i encodes the data (using a simple uniformquantizer with qi bits) prior to information exchange.
1 We also propose to establish a link between any two SNs i and j based on the (known) SNRat node j , i.e.
if SNRij < Υ, eij = eji = 0
if SNRij ≥ Υ, eij = eji = 1.
}
2 Υ is a SNR threshold parameter and SNRij defined as:
SNRij =pt
ijh2ij
ζ0dγij
.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 17 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
Here we propose a scheme, where SN i encodes the data (using a simple uniformquantizer with qi bits) prior to information exchange.
1 We also propose to establish a link between any two SNs i and j based on the (known) SNRat node j , i.e.
if SNRij < Υ, eij = eji = 0
if SNRij ≥ Υ, eij = eji = 1.
}
2 Υ is a SNR threshold parameter and SNRij defined as:
SNRij =pt
ijh2ij
ζ0dγij
.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 17 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
We propose to quantize with qi bits at SN i before transmitting to SN j :
qi ≤1
2log2 (1 + Υ) bits/sample
A large Υ means:
1 Fewer communication links and so slower information diffusion across the network.2 An increase in the number of bits that each SN can transmit to its neighbors.
A small Υ means:
1 Establishes a more connected graph and dictates a faster information diffusion across thenetwork.
2 Allows less transmission bits resulting in an increase in the quantization noise variance.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 18 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
We propose to quantize with qi bits at SN i before transmitting to SN j :
qi ≤1
2log2 (1 + Υ) bits/sample
A large Υ means:1 Fewer communication links and so slower information diffusion across the network.
2 An increase in the number of bits that each SN can transmit to its neighbors.
A small Υ means:
1 Establishes a more connected graph and dictates a faster information diffusion across thenetwork.
2 Allows less transmission bits resulting in an increase in the quantization noise variance.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 18 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
We propose to quantize with qi bits at SN i before transmitting to SN j :
qi ≤1
2log2 (1 + Υ) bits/sample
A large Υ means:1 Fewer communication links and so slower information diffusion across the network.2 An increase in the number of bits that each SN can transmit to its neighbors.
A small Υ means:
1 Establishes a more connected graph and dictates a faster information diffusion across thenetwork.
2 Allows less transmission bits resulting in an increase in the quantization noise variance.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 18 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
We propose to quantize with qi bits at SN i before transmitting to SN j :
qi ≤1
2log2 (1 + Υ) bits/sample
A large Υ means:1 Fewer communication links and so slower information diffusion across the network.2 An increase in the number of bits that each SN can transmit to its neighbors.
A small Υ means:1 Establishes a more connected graph and dictates a faster information diffusion across the
network.
2 Allows less transmission bits resulting in an increase in the quantization noise variance.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 18 / 42
4. Quantized Distributed Soft Decision Fusion Rule
Proposition
We propose to quantize with qi bits at SN i before transmitting to SN j :
qi ≤1
2log2 (1 + Υ) bits/sample
A large Υ means:1 Fewer communication links and so slower information diffusion across the network.2 An increase in the number of bits that each SN can transmit to its neighbors.
A small Υ means:1 Establishes a more connected graph and dictates a faster information diffusion across the
network.2 Allows less transmission bits resulting in an increase in the quantization noise variance.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 18 / 42
4. Simulation Results 1/6
0 10 20 30 40 50 60 70 800
0.5
1
(
Norm
.E# P T
$Proposed two-step
Conventional cons.
0 10 20 30 40 50 60 70 800.68
0.7
0.72
0.74
(
P$ d
0 10 20 30 40 50 60 70 800
0.5
1
(
;
Figure 9: Normalized average power consumption (E[PT
]), achievable8 probability of detection (P∗
d ) and the average
communication link density (ρ) versus Υ, with σ2eh
= 0, decision fusion in (5.4.16), Pgfa = 0.1, U = 3, N = 20, M = 17
and with αi (scaled by M).Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 19 / 42
4. Simulation Results 2/6
0 0.2 0.4 0.6 0.8 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Global prob. of false alarm, P gfa
Glo
balp
rob.
ofde
tect
ion,
Pg d
Centralized optimum linear rule (5.3.12)Proposed weighted two-step, K1 = 100Proposed weighted two-step, K1 = 200Proposed weighted two-step, K1 = 350Proposed weighted two-step, K1 = 500Proposed weighted two-step, K1 = 800Centralized LRT-based [36]
K1 = 350K1 = 200
K1 = 100
K1 = 800K1 = 500
Upper bound
Figure 10: Averaged (over 500 h2ij realizations) ROC for the proposed two-step weighted algorithm with decision fusion
in (40), U = 3, N = 20, M = 17, K2 = 3, Υ = 30, σ2eh
= 0 and with αi (scaled by M) in (5.3.9).
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 20 / 42
4. Simulation Results 3/6
0 0.2 0.4 0.6 0.8 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Global prob. of false alarm, P gfa
Glo
balp
rob.
ofde
tect
ion,
Pg d
Proposed weighted two-step, K1 = 50Proposed weighted two-step, K1 = 150Proposed weighted two-step, K1 = 300Centralized LRT-based [36]Centralized optimum linear rule (5.3.12)
K1 = 300
K1 = 150
Upper bound
K1 = 50
Figure 11: Averaged (over 500 h2ij realizations) ROC against first step iterations number (K1), with decision fusion in
(41), K2 = 2, U = 3, N = 20, M = 17, Υ = 10, σ2eh
= 0 and with αi (scaled by M) in (5.3.9).
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 21 / 42
4. Simulation Results 4/6
-25 -20 -15 -10 -5 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
a
(dB)
Glo
balpro
b.
of
dete
ction,
Pg d
Centr. LRT-based in [36]Centr. opt linear rule (5.3.12)Proposed two-step with (5.4.16)Proposed two-step with (5.4.15)
LRT-based
Centr opt linear rule
Proposed with (5.4.15)
Proposed with (5.4.16)
-20 -15 -10 -5 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
a
(dB)
Glo
balpro
b.
of
dete
ction,
Pg d
Centr. LRT-based in [36]Proposed two-step with (5.4.16)Proposed two-step with (5.4.16)Proposed two-step with (5.4.16)
<2eh
= 4
<2eh
= 1<2
eh= 0
Figure 12: Averaged (over 500 h2ij realizations) probability of detection (Pg
d ) against the signal to noise ratio(ξa) with Pg
fa = 0.1, U = 3, N = 20, M = 17, K1 = 320, Υ = 20, ξi = ξ,∀i in (4) and with αi (scaled by M) in(5.3.9): (left) ideal, σ2
eh= 0; (right) non-ideal, σ2
eh6= 0.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 22 / 42
4. Simulation Results 5/6
0 0.2 0.4 0.6 0.8 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Global prob. of false alarm, P gfa
Glo
balp
rob.
ofde
tect
ion,
Pg d
Proposed weighted two-step with (5.4.15)
Unquantized eq. comb. (,i = 1) in (5.3.14)
Proposed eq. comb. (,i = 1) two-step with (5.4.16)
Proposed eq. comb. (,i = 1) two-step with (5.4.15)
Proposed weighted two-step with (5.4.16)
Centr. opt. linear rule (5.3.12)
K1 = 400
K1 = 320
Centr. opt.
K1 = 320
Centr. eq. comb.
Figure 13: Averaged (over 500 h2ij realizations) ROC for the proposed (quantized) two-step weighted fusion rule
with U = 3, N = 20, Υ = 20, M = 17 and with αi (scaled by M) in (5.3.9).Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 23 / 42
4. Simulation Results 6/6
-14 -12 -10 -8 -6 -4 -20
0.2
0.4
0.6
0.8
1
9a (dB)
Glo
balp
rob
ofde
tect
ion,
Pg d
Unquantized eq comb in (5.3.14)Proposed eq comb two-step, K1 = 10Proposed eq omb two-step, K1 = 20SN 3 eq comb -rst step, K1 = 10
K1 = 20
K1 = 10
Centralized detector
K1 = 10
Figure 14: Probability of detection (Pgd ) versus the signal to noise ratio (ξa) for M = 13, Υ = 72, U = 2, N = 20,
Pgfa = 0.1 and ξi = ξ.∀i in (3.2.4) and αi = 1, ∀i in (5.4.4). The topology used is given in right of Fig. 5.5.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 24 / 42
Overview
1 Introduction
2 Optimal Quantization and Power Allocation
3 Centralized Quantized Fusion Rules
4 Distributed Two-Step Quantized Fusion Rules
5 Sensor Detection in the Presence of Falsified Observations
6 Summary
7 Key Conclusions
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 25 / 42
5. Sensor Detection in the Presence of Falsified Observations
Motivation1 Geographically dispersed to cover large areas, the SNs are constrained in both bandwidth
and power. Usually unattended and this makes them vulnerable to different attacks.
2 The overall detection performance strongly depends on the reliability of these SNs in thenetwork.
3 While fusing the data received by the spatially deployed SNs allows the FC to make areliable decision, it is possible that one or more SNs (compromised by an attacker)deliberately falsify their local observations.
Contributions
1 The problem of centralized detection in the presence of compromised SNs is investigated.
2 Attacker-based and FC-based parameter optimization are considered and some expressionshave been derived.
3 A reputation based scheme to identify the compromised SNs in the network and controltheir influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 26 / 42
5. Sensor Detection in the Presence of Falsified Observations
Motivation1 Geographically dispersed to cover large areas, the SNs are constrained in both bandwidth
and power. Usually unattended and this makes them vulnerable to different attacks.
2 The overall detection performance strongly depends on the reliability of these SNs in thenetwork.
3 While fusing the data received by the spatially deployed SNs allows the FC to make areliable decision, it is possible that one or more SNs (compromised by an attacker)deliberately falsify their local observations.
Contributions
1 The problem of centralized detection in the presence of compromised SNs is investigated.
2 Attacker-based and FC-based parameter optimization are considered and some expressionshave been derived.
3 A reputation based scheme to identify the compromised SNs in the network and controltheir influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 26 / 42
5. Sensor Detection in the Presence of Falsified Observations
Motivation1 Geographically dispersed to cover large areas, the SNs are constrained in both bandwidth
and power. Usually unattended and this makes them vulnerable to different attacks.
2 The overall detection performance strongly depends on the reliability of these SNs in thenetwork.
3 While fusing the data received by the spatially deployed SNs allows the FC to make areliable decision, it is possible that one or more SNs (compromised by an attacker)deliberately falsify their local observations.
Contributions
1 The problem of centralized detection in the presence of compromised SNs is investigated.
2 Attacker-based and FC-based parameter optimization are considered and some expressionshave been derived.
3 A reputation based scheme to identify the compromised SNs in the network and controltheir influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 26 / 42
5. Sensor Detection in the Presence of Falsified Observations
Motivation1 Geographically dispersed to cover large areas, the SNs are constrained in both bandwidth
and power. Usually unattended and this makes them vulnerable to different attacks.
2 The overall detection performance strongly depends on the reliability of these SNs in thenetwork.
3 While fusing the data received by the spatially deployed SNs allows the FC to make areliable decision, it is possible that one or more SNs (compromised by an attacker)deliberately falsify their local observations.
Contributions1 The problem of centralized detection in the presence of compromised SNs is investigated.
2 Attacker-based and FC-based parameter optimization are considered and some expressionshave been derived.
3 A reputation based scheme to identify the compromised SNs in the network and controltheir influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 26 / 42
5. Sensor Detection in the Presence of Falsified Observations
Motivation1 Geographically dispersed to cover large areas, the SNs are constrained in both bandwidth
and power. Usually unattended and this makes them vulnerable to different attacks.
2 The overall detection performance strongly depends on the reliability of these SNs in thenetwork.
3 While fusing the data received by the spatially deployed SNs allows the FC to make areliable decision, it is possible that one or more SNs (compromised by an attacker)deliberately falsify their local observations.
Contributions1 The problem of centralized detection in the presence of compromised SNs is investigated.
2 Attacker-based and FC-based parameter optimization are considered and some expressionshave been derived.
3 A reputation based scheme to identify the compromised SNs in the network and controltheir influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 26 / 42
5. Sensor Detection in the Presence of Falsified Observations
Motivation1 Geographically dispersed to cover large areas, the SNs are constrained in both bandwidth
and power. Usually unattended and this makes them vulnerable to different attacks.
2 The overall detection performance strongly depends on the reliability of these SNs in thenetwork.
3 While fusing the data received by the spatially deployed SNs allows the FC to make areliable decision, it is possible that one or more SNs (compromised by an attacker)deliberately falsify their local observations.
Contributions1 The problem of centralized detection in the presence of compromised SNs is investigated.
2 Attacker-based and FC-based parameter optimization are considered and some expressionshave been derived.
3 A reputation based scheme to identify the compromised SNs in the network and controltheir influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 26 / 42
5. Sensor Detection in the Presence of Falsified Observations
Communication Architecture
SN3
SN2
SN5
SN1
SN4
SN6
Target
Attacker
Fusion Center
Tf =M∑i=1
αiTqi
T fal3
T2
T fal5
T4
T6
T1
T q3
T q2
T q5
T q4
T q1
T q6
Figure 15: Under attack communication architecture between peripheral SNs and the FC. While the honest SNs teststatistics remain unchanged, the compromised SNs falsify their test statistics before transmitting to the FC.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 27 / 42
5. Simulation Results 1/4
1 2 3 4 5 6 7 8 9 10 11 120
1
2
SNs
h i2
1 2 3 4 5 6 7 8 9 10 11 120
10
20
SNs
po i
1 2 3 4 5 6 7 8 9 10 11 120
5
SNs
L i
C=5C=0.5C=0
Figure 16: SN optimal transmit power (poi ) and channel bit allocation (Li ) with Pt = 60, U = 3,
ξa = −10.5 dB, N = 20, β = 0.1 and σ2eh
= 0.Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 28 / 42
5. Simulation Results 2/4
0 0.2 0.4 0.6 0.8 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Probability of false alarm, Pfa
Prob
abili
tyof
dete
ction
,Pd
AF in [35]
opt. in (6.2.22), - = 0:1
OAFBB, - = 0:1
WAFBB, - = 0:1
opt. in (6.2.22), - = 0:5
OAFBB, - = 0:5
WAFBB, - = 0:5
OAFBB, - = 1
WAFBB, - = 1
=0.1
=0.5
=1
Figure 17: Probability of detection (Pd ) versus probability of false alarm (Pfa) with U = 3, Pt = 60,M = 12, N = 20, Ci = 0.9,∀i and σ2
eh= 0.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 29 / 42
5. Simulation Results 3/4
0 0.2 0.4 0.6 0.8 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Probability of false alarm, Pfa
Prob
abili
tyof
dete
ctio
n,P d
C=0C=0.2C=0.4C=0.45C=0.6C=0.9C=1.1C=1.4
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0.86
0.88
0.9
0.92
0.94
0.96
0.98
C=1.4 C=1.1
C=0.9
C=0.2
C=0.45
C=0
C=0.6
Nash Equilibrium, C=0.4
Figure 18: Probability of detection (Pd ) versus probability of false alarm (Pfa), with U = 3, ξa = −10.5dB, Pt = 60, M = 12, N = 20, β = 0.2, σ2
eh= 0 and with optimum weights in (6.2.22).
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 30 / 42
5. Simulation Results 4/4
0 2 4 6 8 10-1
-0.5
0
0.5
1
1.5
2
2.5
3
Attacker strength, C
Modi-
edde.
ection
coe/c
ient,
~ d2Optimum , in (6.2.22)Non-optimum ,First derivative
Figure 19: Modified deflection coefficient (d2) versus the attacker strength (C) with U = 3, ξa = −10 dB,si = 0.1, ∀i , Pt = 60, M = 12, N = 20, β = 0.1 and σ2
eh= 0.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 31 / 42
5. A Secure Sub-optimum Detection Scheme in Under-Attack WSNs
SN3
SN2
SN5
SN1
SN4
SN6
Target
Attacker
Fusion Center
Tf =M∑i=1
αi Ii
I C3
I2
I C5
I4
I6
I1
I3
I2
I5
I4
I1
I6
Figure 20: Under attack schematic communication architecture between peripheral SNs and the fusion center (FC).While the i th (i = {1, 2, 4, 6}) honest SN indicator (test statistic) remains unchanged (i.e., Ii = Ii ), the j th (j = {3, 5})compromised SN falsify its indicator (test statistic) as in (6.3.7) before transmitting to the FC.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 32 / 42
5. A Secure Sub-optimum Detection Scheme in Under-Attack WSNs
FC Optimum Weighting
αiopt =
(1− β
)(pi
d − pifa
)+ β
(pi ,C
fa − pi ,Cd
)(2P fal
C − 1)(
1−β)(pi
d
(1−pi
d
))+β(Pflip
C + pi ,Cd
(1−2Pflip
C
))(1−Pflip
C + pi ,Cd
(2Pflip
C −1)) . (1)
Depends upon the local pifa and the pi
d as well as on the β (fraction of compromised SNs) andthe probability of flipping the local decisions by the attacker. The FC cannot implement theoptimum weight combining fusion ruleAttacker Flipping Probability Optimisation
Lemma 6.3.2: The optimum flipping probability(Pflip
C ,opt
)which minimizes the modified
deflection coefficient is:
PflipC ,opt =
β − 1
2β
( M∑i=1
αi
(pi
d − pifa
)M∑
i=1αi
(pi ,C
fa − pi ,Cd
))
+1
2(2)
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 33 / 42
5. Simulation Results 1/6
5 10 15 200
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
FC threshold, Λf
Repu
tation
metr
ic, r i
SN9SN10SN11SN12SN13SN15
5 10 15 200
0.05
0.1
0.15
0.2
0.25
0.3
0.35
FC threshold, Λf
Repu
tation
metr
ic, r i
SN21SN22SN23SN24SN29SN30SN32SN35SN39
Figure 21: The reliability metric (ri ) versus the FC detection threshold (Λf ) against the SNs with M = 40, N = 20,
β = 0.5, PflipC = 1 and K = 150.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 34 / 42
5. Simulation Results 2/6
0 5 10 15 200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
FC detection threshold, f
Prob
.ofd
et.t
heco
mpr
omise
dSN
37,P
37,t
rue
d
K=5
K=10
K=15
K=20
K=30
K=100
K=200
Figure 22: Probability that the (compromised) SN 37 has been truly detected (P37,trued ) versus the FC detection
threshold (Λf ) with M = 40, N = 20, β = 0.5, PflipC = 1 and δ = 0.009.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 35 / 42
5. Simulation Results 3/6
5 10 15 20 25 30 35 40 45 500
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Time window length, K
Aver.
prob
.of
det.
(mis-det.)
,Ptrue
d(P
false
d)
P trued , Λf = 5
Pf alse
d , Λf = 5
P trued , Λf = 13
Pf alse
d , Λf = 13
P trued , Λf = 13
Pf alse
d , Λf = 13
β=0.5
β=0.25
β=0.10
Figure 23: Average compromised SNs detection probability and honest SNs mis-detection probability versus the time
window length (K) and against β with M = 40, N = 20, PflipC = 1 and δ = 0.009.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 36 / 42
5. Simulation Results 4/6
1 2 3 4 5 6 7 8 9 100.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
Time window length, K
Prob
.ofd
et.m
inus
prob
.off
alse
alar
m,
P d!
P fa
Scheme in [73], $f = 7Scheme in [73], $f = 9Proposed, $f = 7Proposed, $f = 9
f=9
f=7
Figure 24: The Pd − Pfa metric versus the time window length (K) against the FC detection threshold (Λf ) with
M = 40, N = 20, β = 0.25, PflipC = 0.2, δ = 0.95 and µ = 10.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 37 / 42
5. Simulation Results 5/6
0 0.2 0.4 0.6 0.8 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Prob
.of
dete
ctio
n,P d
Prob. of false alarm, Pfa
Upper Bound
Equal combining in [35]
Perfect SNs iden. and tot. removal
Opt. weights (6.3.24), perfect SNs iden.
Proposed, 7 = 15, / = 0:09
Proposed, 7 = 15, / = 0:12
Proposed, 7 = 55, / = 0:009
,i = ,AFi in (6.3.10), no iden. scheme
Proposed, 7 = 35, / = 0:009
Proposed
No ident. scheme
Figure 25: Probability of detection (Pd ) versus probability of false alarm (Pfa) with M = 40, N = 20, β = 0.5, PflipC = 1
and K = 5.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 38 / 42
5. Simulation Results 6/6
0 0.2 0.4 0.6 0.8 10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Prob. of false alarm, Pfa
Prob
.of
dete
ctio
n,P d Upper Bound
Opt. weights (6.3.24), perf. SNs iden.
Equal combining in [35], - = 0
Proposed, / = 0:009, K=80, 7 = 6
Proposed, / = 0:009, K=40, 7 = 6
Proposed, / = 0:009, K=5, 7 = 10
Proposed, / = 0:009, K=5, 7 = 14
Proposed, / = 0:009, K=5, 7 = 1
With ,i = ,AFi in (6.3.10), no iden.
Scheme in [73], / = 1, K = 5
Equal combining, - = 0:25
Proposed
Eq. comb.
No iden. scheme
Scheme in [73]
Figure 26: Probability of detection (Pd ) versus probability of false alarm (Pfa) against δ and µ with M = 40, N = 20,
β = 0.25, and PflipC = 1.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 39 / 42
Summary
We derive the optimum fusion rule and then analyze sub-optimum fusion rules that arerealizable and easily implemented in practical WSN deployment scenarios. The effect offading channels on detection performance is minimized by solving the resource allocationproblem.
A two-step consensus-based approach with weight combining quantized test statisticsexchange is proposed. We relate the communication topology with the number of bits tobe shared among SNs. It turns out that there is an optimum topology that maximizes thedetection performance.
Centralized detection in the presence of compromised SNs is also investigated. Attackerand FC based parameter optimization are considered and some expressions have beenderived. A reputation based scheme to identify the compromised SNs in the network andcontrol their influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 40 / 42
Summary
We derive the optimum fusion rule and then analyze sub-optimum fusion rules that arerealizable and easily implemented in practical WSN deployment scenarios. The effect offading channels on detection performance is minimized by solving the resource allocationproblem.
A two-step consensus-based approach with weight combining quantized test statisticsexchange is proposed. We relate the communication topology with the number of bits tobe shared among SNs. It turns out that there is an optimum topology that maximizes thedetection performance.
Centralized detection in the presence of compromised SNs is also investigated. Attackerand FC based parameter optimization are considered and some expressions have beenderived. A reputation based scheme to identify the compromised SNs in the network andcontrol their influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 40 / 42
Summary
We derive the optimum fusion rule and then analyze sub-optimum fusion rules that arerealizable and easily implemented in practical WSN deployment scenarios. The effect offading channels on detection performance is minimized by solving the resource allocationproblem.
A two-step consensus-based approach with weight combining quantized test statisticsexchange is proposed. We relate the communication topology with the number of bits tobe shared among SNs. It turns out that there is an optimum topology that maximizes thedetection performance.
Centralized detection in the presence of compromised SNs is also investigated. Attackerand FC based parameter optimization are considered and some expressions have beenderived. A reputation based scheme to identify the compromised SNs in the network andcontrol their influence to the global FC decision is also proposed.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 40 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms
=⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing =⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision =⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach =⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms =⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing =⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision =⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach =⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms =⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing
=⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision =⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach =⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms =⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing =⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision =⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach =⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms =⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing =⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision
=⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach =⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms =⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing =⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision =⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach =⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms =⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing =⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision =⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach
=⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Key Conclusions
Shown that spatially distributed SNs across the field can offer a reliable operation for eventdetection applications. The system detection performance and the WSN’s operating lifetime canbe further improved by means of resource allocations, optimisation and signal processingalgorithms =⇒ complexity to be kept as simple as possible.
The data fusion problem: we derive the optimal fusion rules (i.e., for attack-free and under-attackWSN scenarios) and have shown that these fusion rules are not implementable in practice andrequire complex local signal processing =⇒ Derive sub-optimum but simple fusion rules (requiringsimple hardware) that offer reliable and good detection performance.
A better but more complex approach is to possibly identify these compromised SNs and controltheir influence on the FC decision =⇒ Offers an improved detection performance but requiresobserving the SN’s local reports for a period of time. A larger observation time period (K) maylead to a large detection delay that is critical for most of the event detection applications.
We have addressed the fully distributed detection problem and proposed signal processingalgorithms for such an approach =⇒ Very attractive from both the signal processing perspectiveand the communication point of view.
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 41 / 42
Questions/Comments
Edmond Nurellari (University of Leeds) Distributed Detection and Estimation in WSNs June 6, 2017 42 / 42