Dynamic Threshold Selection Approach in Voting Rule for Detection of Primary User Emulation Attack

Abstract

Cognitive radio (CR) technology presents a mechanism for efficient spectrum usage. Spectrum sensing is an essential CR function which includes an intelligent signal processing algorithm to identify the vacant frequency bands. Cooperative spectrum sensing (CSS) has been widely adopted to improve the performance of CR networks. Unfortunately, CR networks are vulnerable to security threats. In this study, we propose an optimal threshold selection approach to address one of the most important attacks called primary user emulation attack (PUEA). In PUEA, a malicious attacker mimics some important primary signal features and deceives CR sensors to prevent them from accessing the available channels. In this study, we assume a malicious PUEA which is relatively located near the potential user (PU) transmitter and senses the spectrum and accurately detects the vacant frequency bands to transmit its fake signal. We estimate attack strength and then apply the K-out-of-N rule to obtain an optimum and dynamic threshold K, minimizing the global error probability. Here, the attack strength is defined as the ratio of the average transmission power of the PUEA to the average power of the PU. The achieved simulation results indicate that the performance of the suggested method is satisfactory in detecting the malicious PUEA compared with conventional methods.