A Different Approach for Sensing Disruption

Spectrum sensing vulnerabilities in cognitive radio networks can significantly degrade performance. Most disruption attacks in the current literature involve spoofing of the free bands used for sensing by making them appear busy. In this study, we proposed a different approach for sensing disruptions. We examined the optimal strategy for an intelligent adversary with a given power to flip busy bands and make them appear free. The mechanism of sensing disruption was established by contaminating the noise power measurements. This is illustrated by a two-step sensing scheme in which energy detection, in conjunction with noise power estimation, is used by secondary users. We show that to flip busy bands, the optimal strategy for sensing link disruptions is equal-power, and partial-band flipping. We demonstrated that the maximum average number of missed detections can be derived under a constraint on power of the adversary. Through analytical and numerical results, we demonstrated the effectiveness of our approach in terms of the impact of disruption attacks on spectrum sensing.