Misbehavior Detection and Identification in DF Multi-relay VANETs

In this paper, we study the multi-relay misbehavior detection and identification problem in decode-and-forward (DF) vehicular ad hoc networks (VANETs). To simplify detector design, we present a novel metric, hypothesis variance ratio (HVR), by which the closed-form expressions of the false-alarm rate, miss-detection rate, weighted detection error probability (WDEP) and the optimal threshold to minimize the WDEP are derived. Furthermore, the monotonicity of the above metrics with respect to the HVR is proved. Based on the above theories and the sparsity of relay misbehaviors, we develop a pre-detection scheme to reduce computation latency. The maximum HVR based sufficient statistic is constructed to achieve high latency reduction rates. It is demonstrated that the proposed scheme outperforms the average-power-Neyman-Pearson scheme in terms of the WDEP.