摘要：As the effective supplement to traditional cryptography security, trust management plays a key role for detecting malicious behaviors in wireless sensor networks. In this paper, we consider a sensor network wherein nodes have fixed number of states and probabilities of state transitions can be specified. We introduce Markov chain model and Fokker-Planck equation for evaluating the short-term trust value of nodes. In order to avoid the drawbacks of existing long-term trust evaluation models, we propose a new evaluation model with adaptive forgetting factors, where step functions are introduced to calculate the current forgetting factor. Simulation results show that the proposed trust management scheme can effectively detect malicious nodes in wireless sensor networks, and significantly improve the packet delivery ratio compared to the counterparts
