Analysis and Detection of Cyber-physical Attacks in Distributed Sensor Networks

This paper analyzes and detects the adverse effects of cyber-physical attacks on distributed Kalman filters in sensor networks. To this end, we first evaluate the effect of adversaries on the performance of the sensor network in the distributed state estimation problem. More specifically, we consider attacks on both the sensors and the wireless communication channels and show that how an attacker affects the state estimation error covariance recursion and, consequently, the network performance. We then introduce novel Kullback-Liebler divergence based detectors to capture attacks by using the innovation sequence of each sensor and the innovation sequences that it estimates for its neighbors. Simulation result validates the effectiveness of presented work.