Reputation-Based Self-Differential Sequential Mechanism for Collaborative Spectrum Sensing Against Byzantine Attack in Cognitive Wireless Sensor Networks

In order to meet the increasing frequency demand for sensors and their related applications, cognitive radio (CR) technology has been integrated into wireless sensor networks, detecting available spectrum resources through collaborative spectrum sensing (CSS) among multiple sensors and avoiding harmful interference to the primary user. However, some malicious sensor nodes (MSNs) may also take advantage of collaborative opportunities to launch Byzantine attack, reducing the performance and efficiency of CSS. In order to suppress the negative impact of MSNs, this letter proposes a reputation-based self-differential sequential mechanism (R-SDSM) to defend against Byzantine attack. First, sensor nodes with high reputation value are prioritized to participate in CSS and complete the data fusion with more appropriate weight allocation. Furthermore, a self-differential sequential mechanism is proposed to reduce the reporting decisions required for the fusion center. Finally, numerical simulation results demonstrate that in contrast to other data fusion rules, the proposed R-SDSM provides higher detection accuracy and fewer reporting decisions.