Research Of Trust Management In Wireless Sensor Networks

Wireless sensor networks are composed of a large number of low-cost micro sensor nodes which has the ability of communication and awareness deployed in the monitoring area. It's a multi-hop autonomous network system which forms by wireless communication methods.As a system which used to access information between people and between human and nature, it has very broad application prospects and developing potential, especially in the severe and unattended environment.Wireless sensor network is an autonomous network which has no infrastructure and communicates by wireless links.Each sensor node is very small and limited in its memory, computation, battery life and communication capabilities.They are susceptible to a variety of attacks, for example node capture, denial of services,eavesdropping and so on. Therefore, constructing a complete trust model for wireless sensor networks has a strong practical significance for its applications.This is also the purpose of this paper. Now our main work is as follows:Firstly, we studied the cloud-based trust model.CBTM took the uncertainty, fuzziness and randomness of trust between nodes into account completely, and quantified the trust between nodes correctly.It was more accurate and comprehensive than RFSN.But this model didn't take the timeliness of trust relationship between nodes into account, and the trust combination which adopted the method of averaging directly couldn't resistance bad-mouthing attacks or ballot-stuffing attacks which were initiated by malicious nodes effectively.Aiming at these problems,a new trust evaluation model for wireless sensor networks based on trust-cloud was proposed. In order to improve the ability of resistance malicious attacks,this model distributed the weights of the recent trust-cloud and the historical trust-cloud according to the recent behavior of nodes,and amended the weights of the direct trust-cloud and the recommended trust-clouds using the similarity degree respectively.Secondly, we studied the factors which could impact the trust of wireless sensor networks,and proposed a trust model for wireless sensor networks based on multi-angle combination of the communication, data and energy.This model took all the factors into account to accelerate the convergence of trust in the network and exclude the unreliable nodes and unreliable data as early as possible. Finally, we verified the validity of their abilities to resistance malicious attacks,exclude unreliable, and identify malicious nodes by simulating the two trust models using MATLAB.