Malicious Nodes Detection Scheme Based On Dynamic Trust Clouds for Wireless Sensor Networks

The randomness, ambiguity and some other uncertainties of trust relationships in Wireless Sensor Networks (WSNs) make existing trust management methods often unsatisfactory in terms of accuracy. This paper proposes a trust evaluation method based on cloud model for malicious node detection. The conversion between qualitative and quantitative sensor node trust degree is achieved. Firstly, nodes cooperate with each other to establish a standard cloud template for malicious nodes and a standard cloud template for normal nodes, so that malicious nodes have a qualitative description to be either malicious or normal. Secondly, the trust cloud template obtained during the interactions is matched against the previous standard templates to achieve the detection of malicious nodes. Simulation results demonstrate that the proposed method greatly improves the accuracy of malicious nodes detection.