Detection of Abnormal Power Emission in UAV Communication Networks

Abstract

Abnormal power emission poses serious interference and security threats to unmanned aerial vehicle (UAV) communication networks. To tackle this problem, in this letter, we first develop a cloud-based drone surveillance framework, where a closed-loop cognitive control cycle is formed with power requirement and allocation information exchanged between a UAV network and a management cloud, and allocation information and abnormal power emission exchanged between the cloud and a surveillance center. Then, the detection problem of abnormal power emission is mathematically formulated as a ternary-hypothesis test. To achieve the tradeoff between false alarm and missed detection, a generalized Neyman-Pearson test criterion is proposed, where the detection probability is maximized under two false-alarm constraints. Following the criterion, the test rules are derived for local detection and cooperative detection, where local decision regions and the global decision threshold are optimized, respectively. Finally, the simulation results are provided to verify the proposed scheme.