基于节点相似性的WSNs故障检测方法研究

Research on WSNs fault detection method based on node similarity

针对目前多数无线传感器网络分布式故障检测的算法都以假设故障节点数据为离群值为基础,存在局限性的问题。提出一种基于节点相似度比较的无线传感器网络故障检测方法,簇头节点根据簇内节点数据的时空相关性,进行节点相似性度量,实时调整节点可信水平,并采用最优函数计算出当前实验的最优阈值(0.8)进行故障节点的判断。通过仿真实验证明:针对不同的故障模型,算法保持了良好的故障检测能力,一定程度上解决通用性问题。

Aiming at present distributed fault detection algorithms for wireless sensor networks （WSNs）have assume that the fault node are based on outliers data,it have limitations.Present a method based on similarity comparison of nodes of WSNs fault detection,according to correlation of time and space of node data within the cluster,the cluster head nodes measure the similarity among cluster nodes,and adjust node confidence level real-time,calculate the optimal threshold value which is 0.8 by using optimal function in current experiments to judge fault node.Through simulation experiments show that aiming at different fault model,the algorithm keep in good ability of fault detection,and solve the problem of generality in a certain extent.