In order to solve the problem that the existing trust models could not solve malicious nodes in wireless sensor networks effectively and the problem that counter-intuitive results may appear when combining evidence, a detection scheme based on improved D-S evidence theory was proposed. The inherent deficiency of the D-S evidence theory was analyzed, a new variable was defined to describe the difference among trust evidence, and the original D-S evidence theory combination rule was modified to obtain more accurate results. Based on those improvements, a detection scheme to identify and isolate misbehaviors nodes in WSNs was proposed. Moreover, simulations were carried out to verify the effectiveness of the proposed scheme. Simulation results showed that this model could find and isolate malicious nodes quickly and accurately, suppress nodes collusion and improve network performance. Compared with other existing trust models, this model has more security, robustness and accuracy.
