Physical Layer Security in Wireless Sensor Networks

Abstract

Due to the broad application of wireless sensor networks (WSNs), the security of these systems have become a necessity. This thesis focuses on the physical layer security of WSNs using chaotic Direct Sequence Spread Spectrum (chaotic-DSSS). The thesis begins with investigating the overall security threats to WSNs, whereby the possible attacks on different Open Systems Interconnection (OSI) layers have been reviewed and analyzed, showing their vulnerability to various security attacks due to hardware limitation of WSNs. One of the primary attacks against WSNs is jamming, the effect of which on the physical layer of WSNs is investigated in this thesis. The WSNs in this study use chaotic-DSSS modulation, as Chaotic-DSSS provides secrecy and jamming resistance. The jamming resistance performance of chaotic-DSSS under the perfect chip synchronization is evaluated. Chaotic-DSSS is extremely dependant on chip synchronization. The synchronization performance against different jammers is also investigated based on the probabilities of detection and false alarm. It is shown that chip synchronization in chaotic-DSSS is vulnerable to jamming attacks. The effect of the probabilities of detection and false alarm on bit error rate is investigated by proposing the probability of successful synchronization. The results demonstrate the significant effects of jammer on the probability of bit error of chaotic-DSSS in the case of imperfect chip synchronization. Based on this vulnerability, this thesis also proposes a new security attack called Syncim. In this attack, the adversary interferes with the synchronization pilot signals of the legitimate transmitter. Then, it sends its fake synchronization pilot signals and makes the receiver (victim) sync with the adversary instead of the legitimated transmitter. The sensitivity of the chaotic-DSSS to the chip synchronization has been an incentive for researchers to investigate different methods to remove the chip synchronization and maintain the secrecy feature of the chaotic-DSSS. Furthermore, the chaotic-DSSS system requires a random number generation for each chip, which makes it complicated to implement in WSNs. This thesis proposes a fix chaotic-DSSS system, and investigates its secrecy performance. Also, a fake user method is proposed to enhance the secrecy of the fix chaotic-DSSS.