Collaborative Spectrum Sensing in the Presence of Byzantine Attacks

Cognitive radio (CR) has emerged as an enabling technology for supporting dynamic spectrum access (DSA). DSA enables the solution of the spectrum scarcity problem by exploiting transmission opportunities in the underutilized spectrum bands of primary users. In collaborative (or distributed) spectrum sensing, multiple CRs collaborate with each other by sharing their inference to make a global inference regarding the availability of the spectrum, i.e., licensed primary channels. This approach has been shown to have various advantages in terms of spectrum utilization and robustness. A key component of collaborative spectrum sensing is the distributed data fusion scheme. However, the distributed nature of these schemes makes them quite vulnerable to adversarial attacks. A potential attack on this scheme is the Byzantine (or data falsification) attack where some of the CRs are reprogrammed by an adversary, and these malicious CRs (also referred to as Byzantines) introduce false sensing information in the fusion process to degrade the spectrum-sensing process. In this chapter, we analyze the performance limits of collaborative spectrum sensing under Byzantine attacks for different practical network architectures. Specifically, we begin the study from an attacker's perspective and discuss the optimal attacking strategies for given attacking resources. Next, we discuss possible countermeasures from a network designer's perspective and present robust spectrum-sensing algorithms. Finally, we discuss some of the main challenges and research opportunities that exist in emerging CR-based sensing networks. This chapter provides comprehensive insights into the use of signal processing methods for robust spectrum sensing in cognitive radio networks.