A secure distributed consensus scheme for wireless sensor networks against data falsification

This paper focuses on the problem of secure distributed consensus to defend covert misbehavior in wireless sensor networks (WSNs). Distributed consensus is a promising method to improve the efficiency and precision of consensus results in WSNs, but it introduces new security issues that malicious nodes may manipulate false sensing data to degrade the sensing result of the whole network. A data falsification attack, i.e., the attacker injects random values into its neighboring nodes at each time-step of consensus process, is considered. This kind of attack cannot be defended against by most of existing detection mechanisms. We present a distributed detection mechanism with adaptive local threshold to isolate the abnormal nodes. A Weighted Averaging-based Consensus Scheme (WACS) is proposed to decrease the negative impact of the attack and make the network converge to a consensus value. It is proved that convergence property can be guaranteed by the relationship between weighted average of the noise and stochastic approximation. Simulation results are presented to show the effectiveness of the proposed secure scheme.