Trust-based anomaly detection in wireless sensor networks

Due to the openness of the wireless media and frequent interactions among sensor nodes, security has been tightly related to the data credibility and network reliability in wireless sensor networks (WSNs). In this paper, based on fuzzy theory and evidence theory, we present a novel trust model to detect anomaly nodes in WSNs. Specifically, by observing nodes' behaviors with multi-dimensional characteristics, the proposed model can be used to efficiently identify malicious nodes while validating the normal operations. Extensive simulations are conducted, and the simulation results demonstrate the proposed trust model can achieve higher detection rate of malicious nodes in comparison with previously reported ones.