Defense Against Byzantine Attack in Cognitive Radio Using Isolation Forest

Cognitive Radios (CRs) have gained significant popularity due to their ability to harness white spaces opportunistically to improve the spectrum utilization of underutilized spectrum. Cognitive Radio uses cooperative spectrum sensing (CSS) to detect free licensed spectrum. In this work, centralized CSS is explored where secondary users (unlicensed users) sense a band that belongs to the primary user and reports their observations to the fusion center (FC), which takes the final decision to access the channel. With the openness of cognitive radio networks (CRN), there exist numerous attacks which can lead to various security threats in CR. One such attack is the Byzantine attack, also known as the spectrum sensing data falsification (SSDF) attack. This paper proposes an isolation forest-based anomaly detection algorithm for detecting attackers in the CRN. Various simulations are carried out to show the effectiveness of the proposed algorithm in detecting malicious users in different attack scenarios.