Dynamic Trust Evaluation Of Wireless Sensor Networks Based On Multi-factor

As a result of the limitation of sensor nodes and the variability of application environment,wireless sensor networks are often vulnerable to various attacks.Traditional cryptographic mechanism is a passive prevention technology to resist external attacks but invalid for insider attacks.As a complementary technology to the cryptographic mechanism,trust mechanism can solve the problems resulted from insider attacks effectively.In view of the singleness of trust evaluation factors,the inconsistence of weight allocation algorithms and network features and the rough granularity of classifying recommendation trusts,this paper put forward a dynamic trust evaluation method based on multiple factors,and designed and implemented a dynamic trust evaluation platform of wireless sensor networks.First,combining with the characteristics of wireless sensor network and trust,we analysed and summarized evaluation factors that may affect the trust value of node and their access.Secondly,we defined their measures.Then,under the background of Hoeffding's Inequality,we put forward a classification standard and a dynamic weight assignment based on interaction times.We combined direct trust value with indirect trust value to evaluate the nodes' trustworthiness comprehensively and according to the evaluation result to judge node trusted or not.Finally,our experiments demonstrated the effectiveness of trust evaluation method in this paper.Trust evaluation method presented in this paper analyzes the influencing factors of the node trust value from the multiple perspectives and makes an improvement of the trust calculation process.Using this method can quickly identify malicious nodes network,improve the network throughput and correct delivery rate so as to reduce the network security risks,enhance the purpose of network security.