Investigations of Radio Behavior and Security Threats in Cognitive Radio Networks

Cognitive radio (CR) has been an active research topic in wireless communications. In this dissertation, radio behavior and potential security threats in cognitive radio networks are investigated in order to successfully deploy CR networks and realize its benefits. We study complex CR misuse issues during its secondary access processes, including misbehavior, cheating, and attack. Different countermeasure methods are also examined and evaluated. A new type of cognitive attack, most active band (MAB) attack, is proposed in this dissertation, where an attacker or a malicious CR node senses/determines the most active band within a multi-band CR network and targets this band through a denial of service (DoS) attack. A countermeasure strategy, coordinated concealment strategy (CCS), is introduced to counter the MAB attack. Our results show that CCS significantly outperforms CR’s inherent capability of signal/interference avoidance under a MAB attack. We also introduce CR’s power control capability in CCS to further improve the countermeasure performance. Further, to mitigate the misuses of CR networks, radio identification or classification is an important research issue. This dissertation presents two innovative radio modeling approaches for cognitive radio monitoring and identification in this dissertation. In radio fingerprinting modeling, we consider radio frequency fingerprinting (RFF) features that due to the radio’s hardware imperfections are more difficult to reproduce for misuse. We implement an RFF based radio modeling approach considering an RF transient waveform. Using machine learning algorithms, we analyze the extracted RFF features to identify radio terminals. A test bed is