Research On Node Trust Model Based On Bayes Improved Method In Wireless Sensor Network

In recent years, wireless sensor network(WSN) has been applied in many fields, intrusion detection system based on wireless sensor networks is still a key technology in security network system. In this paper, the intrusion detection based on node behavior trust model is studied. This paper presents the node trust model based on Bayes Improved method(NTM-IBM) model, and calculates the trust degree of invaded nodes.Firstly, we introduce importance of intrusion detection, the characteristics of intrusion detection systems and causes of false alarms in intrusion detection. The NTM-IBM model and index to evaluate the model are also presented.Secondly, through the analysis of the Bayes method, this paper applies the penalty factor to the Bayes improving method and improves the calculation on the comprehensive trust. Considering the changing topology of wireless sensor networks,nodes resource-constrained characteristics, this paper analyzes the data acquisition software for wireless sensor network, gets a sample of information nodes, and calculates the node trust. It then compares three attacks, the wormhole attack, Hello flooding attack and without attack, in the detection rate and missing rate in the intrusion. The evaluate performances in NTM-IBM model is also given.Then, the application of NS2 simulation software is used for the model. The model can compute nodes trust very well. Comparing with RFSN model and TCM-BE model, the NTM-IBM model not only has the same security, and has a small amount of communications calculation. The fitness of the node’s trust is more higher.Finally, the evaluation model accesses three indexes: the detection rate, the miss rate and energy consumption. Experimental results show that NTM-IBM model has lower energy consumption, high detection rate, low high missed rate and high detection accuracy. The model is a good solution to the invasion of abnormal or malicious nodes.