Trust evaluation method for Wireless Sensor Networks based on behavioral similarity and similarity coefficient

Recently, the security of Wireless Sensor Networks (WSNs) has been taken match attention because WSNs are extensively applied in many applications such as environment monitoring, military operations, critical infrastructure monitoring. Trust factor has been proved to be an efficient strategy that secure the network from compromised and misbehavior nodes. Several trust models have been proposed. However, the majority of them did not consider the uncertainty nature of trust. Therefore, to overcome the uncertainty of trust, we suggest a new trust method using the behavioral similarity between sensor nodes. First, we break the behavior of sensor node into a set of behavioral patterns. Next, we use different threshold values to differentiate between trustworthy and untrustworthy patterns. Using bipartite graph, we represent the sensor nodes and their behavioral patterns as two groups of vertices. Through vertex similarity, the trust value is measured in term of behavioral similarity between the sensor nodes. Finely, we evaluate our scheme using MATLAB as a simulation tool. The result indicates that our scheme achieves high detection rate of malicious nodes and protects the network against attacks.