Generalized Byzantine Attack and Defense in Cooperative Spectrum Sensing for Cognitive Radio Networks

Cognitive radio (CR) is a revolutionary paradigm to solve the spectrum scarcity problem in wireless networks. In cognitive radio networks, cooperative spectrum sensing is regarded as a promising approach method to significantly improve the performance of spectrum sensing, but it can be threatened by Byzantine attack. The existing defense references have focused on how to mitigate the negative effect of Byzantine attack, but with some strong assumptions, such as the attackers are in minority and/or a trusted node exists for data fusion. This observation motivates us to comprehensively analyze the strategies of Byzantine attack and the fusion center (FC) in the absence of these restrictions. To be specific, we consider a generic Byzantine attack model by analyzing sophisticated malicious behaviors, which goes beyond the existing models for its generalization. Under this generalized attack model, we derive the condition which makes the FC blind from malicious perspective. On this basis, the optimal attack strategy to maximize Bayes risk is analyzed, respectively, in the case of the unknown and known fusion rule. Furthermore, we extend our analysis to the scenario where the FC has the knowledge of the attack strategy by an estimation algorithm and adopts the optimal fusion rule. Thus, we also give the closed form expression, in terms of the optimal attack strategy under different scenarios, sequentially. At last, the extensive numerical results are provided to verify our theoretical analyses and proposed estimation algorithm.