Research Of Fault Detection Method For Hierarchical Wireless Sensor Network

With the rapid development of sensor technology, wireless sensor network technology is also widely used in various fields.However, due tothe limited sensor resources, and workingfor a long time, oncesensor nodes failed, the network will become unstable, soit is important to do researchonwireless sensor network fault detection technique. Currently the distributed fault detection methods arewidely used. However, the energy of nodes consume fast, and generate a large amount of redundant calculations, therefore, the study of a low energy consumption of wireless sensor network fault detection method for prolonging network lifetime, making the network stable and reliable operation is very meaningful.In this paper, based on the spatio-temporal correlation of sensor nodes, designed hierarchical wireless sensor network fault detection method--CFD algorithm, it is mainlyfor the error messagewhen collecting or transferring data generated by the node in wireless sensor networks. Firstly, use a hierarchical network topology control technology to cluster the network. Secondly, use the method ofmiddle value sampling to verify the reliability of the cluster head node. And ultimately make use of the characteristics of that sensors’ sensing data have similar spatial, and then referred to the reliable cluster head node to complete fault detection within a cluster. In the process of fault detection,in order to determine the status of member nodes,which only need tocommunicate withtheir cluster head. Compared with the distributed fault detection method, CFD algorithm can avoid a lot of redundant information generated when the neighbor nodes communicate with each other, and from the fundamental solution to solve the problem of high energy consumption nodes. At the same time, the algorithm also set a confidence value for each node, by the fault tolerance mechanisms to achieve instantaneous fault tolerance.From algorithm simulation results we can see thatthe average energy consumption of the network proposed by CFD method than the distributed approach to save nearly twice as much.