无线传感器网络节点故障诊断算法研究  

On Fault Diagnosis Algorithm in Wireless Sensor Network Nodes

作　　者：许红艳 郭红 

机构地区：[1]郑州澍青医学高等专科学校卫生管理系,郑州450064

出　　处：《西南师范大学学报（自然科学版）》2016年第9期126-133,共8页Journal of Southwest China Normal University(Natural Science Edition)

摘　　要：针对无线传感器网络(WSN)中传感数据缺失或偏离真实值从而降低网络运行可靠性的问题,开展了关于WSN中各种节点故障的检测和诊断算法的研究,提出了基于SVDD的无线传感器网络Outlier检测、多尺度主元分析的WSN节点故障诊断算法,并进行了实验验证.实验结果表明本文方法能够有效实现对节点运行状态的监视和诊断,确保感知数据的准确性和完整性,提高网络运行的可靠性.The impact of the unique features of WSN and the extern factors, such as strict resource con-straints ,sensor aging, harsh or atrocious environments and human destruction, will make the sensed measurements deviated from the true values or miss, thus the network reliability. In order to popularize the application of WSN, it has important values of theoretical researches and practical applications to de-velop efficient fault diagnosis algorithms in WSN nodes, which realizes the surveillance and diagnosis of nodes working condition, ensure the accuracy and integrity of the data collected from WSN, and improve the reliability of WSN. The causes and classification methods of WSN fault have been analyzed, and then the fault diagnosis methods in WSN nodes based on support vector machines （SVM） and principal compo-nent analysis（PCA） method been studied in this dissertation. Experimental results show that this research method can effectively monitor and diagnose the running state of the node, and this method ensures the accuracy and integrity of the data, and improves the reliability of the network operation.

关 键 词：无线传感器网络 故障诊断 支持向量数据描述 多尺度主元分析 递归主元分 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]