An Adaptive Deviation-tolerant Secure Scheme for Distributed Cooperative Spectrum Sensing

Distributed collaborative spectrum sensing is a promising method to improve the precision and efficiency of primary user detection in cognitive radio networks. Despite its performance advantages, it introduces new security issues that malicious or selfish nodes may manipulate false sensing data to degrade or even covert the sensing result of the whole network. Existing research often utilizes a threshold to distinguish honest users and malicious ones. However, determining such a threshold is difficult due to the dynamic characteristic of cognitive radio networks, and it is likely to misjudge an honest node with a relatively large deviation to be malicious. In this paper, we propose an Adaptive Deviation-tolerant Secure Scheme (ADS) for distributed collaborative spectrum sensing, which aims to mitigate the misbehaviors of inside malicious nodes and, at the same time, tolerant the large deviation introduced by honest users. ADS achieves the trade off of sensing security and deviation tolerance by assigning a dynamic weight to each sensing node and utilizes an adaptive threshold to minimize the negative effect on honest users. We evaluate the performance of the scheme through both analytical and simulation based study.