Sequential spectrum sensing against random Byzantine attack in cognitive radio networks

Abstract Cognitive radio (CR) technology allows secondary users (SUs) to opportunistically access the licensed spectrum being authorized by the primary users (PUs) and then improves the spectrum utilization. Cooperative spectrum sensing (CSS) is an important component of cognitive radio (CR) technology to detect the PU signal via spatial diversity. However, due to the open facet of CR networks (CRNs), the CSS process may suffer from Byzantine attack where malicious users (MUs) falsify the status information about the PU signal and then cause harmful interference to the PU's communication or selfishly occupy spectrum resources. Otherwise, the cooperative paradigm also results in that multiple SUs need more time to detect the PU signal. The performance and efficiency of CSS will be further degraded due to these two reasons. For this aim, we propose sequential spectrum sensing (SSS) scheme to mitigate the negative impact of Byzantine attack and improve the cooperative efficiency. To this end, encouraged by in‐depth analysis of the blind problem, we formulate a random Byzantine attack model to follow the malicious behaviors from the perspective of MUs. In order to solve the blind problem in the process of the sensing sample fusion, a delivery evaluation mechanism is developed to lay solid foundation for SSS. On this basis, we further evaluate the reputation value (RV) within a sliding window and in order to improve the cooperative efficiency in the reporting stage, we dynamically select the sensing results of the more reliable SUs as the global decision to make a tradeoff between cooperative performance and cooperative efficiency. At last, extensive simulation results show that regardless of the malicious ratio, our proposed SSS only requires 6 samples to provide with 100% detection accuracy in the presence of always attack while also the remarkable performance in the presence of random attack when the malicious ratio exceeds 50%.