Node Diagnosis Method Of Clustering Wireless Sensor Networks

Sensor nodes in the wireless sensor networks have the design of the limitations in the function of the hardware and software, such as memory storage, enery supply. It is easy to cause node failure, reduce or losing the ability to monitor. Therefore,the node fault detection, fault isolation and data recovery are important to improve the reliability and security of the network operation.This paper puts forward effective method of node fault diagnosis in wireless sensor networks, specific content is as follows:Firstly, this paper puts forward a kind of model(LSI-BM) for sensor node fault diagnosis. On the basis of data time correlation of node, this model is established based on vector hybrid product of abnormal data diagnosis criteria and node trust value evaluation method based on bayesian theory. In addition, for the fault node missing data, using the Hausdorff distance between variables on the similarity measure. In order to avoid the adjacent fault node data influence on the final result, we put forward a data recovery method based on weighted node trust value. The simulation results show that LSI-BM model can effectively diagnose the abnormal data and fault node, and data recovery method to realize the effective data to fill.Secondly, we review the significant topology control algorithms to provide insights into how energy efficiency is achieved by design. To improve the efficiency of energy use in large-scale network nodes, we improve the uneven equal radius of cluster nodes in the campaign(EEUC) agreement and cluster heads selection mechanism. Then we put forward an unequal hierarchical clustering protocol(UHC). In addition, for node distribution random number mechanism may cause campaign uncertain problems. We adjust node random number threshold, which is no longer a fixed value. Finally,through computer simulation, compared the UHC agreement and EEUC based network clustering effect, and the simulation results show that UHC agreement effectively reduce cluster-heads node density, prolong the service life of network, and cluster distribution presents certain regularity.