物联网感知层传感节点故障诊断研究  

Research on Sensor Node Fault Diagnosis of Internet of Things Perception Layer

作　　者：梁小晓[1] 曹莉[2] 韦崇岗[2] 乐英高[3] 

机构地区：[1]四川理工学院计算机学院,四川自贡643000 [2]四川理工学院自动化与电子信息学院,四川自贡643000 [3]东南大学仪器科学与工程学院,南京210096

出　　处：《组合机床与自动化加工技术》2015年第3期62-66,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：企业信息化与物联网测控技术四川省高校重点实验室项目(2013WYJ03;2014WYJ04)

摘　　要：针对物联网感知层节点故障诊断问题,提出基于蚁群聚类优化RBF神经网络的WSNs节点故障诊断算法。将从节点硬件模块故障和节点故障率对故障诊断精度的影响两个方面研究WSNs节点故障诊断。将改进的蚁群聚类优化RBF神经网络的初始权值应用到WSNs节点故障诊断研究中。利用蚁群算法并行寻优特性和自适应调整挥发系数特征作为聚类算法来确定RBF神经网络初始权值,同时采用裁剪约简RBF神经网络隐含层、优化网络结构。通过实验结果表明,基于蚁群聚类优化RBF神经网络的WSNs节点故障诊断方法能准确实现感知节点的故障诊断,与其它方法相比具有更高的诊断精度。In viewof Internet of Things networking perception layer node fault diagnosis problem,this article put forward to WSNs node fault diagnosis method of optimize the RBF neural network based on Ant Colony Clustering Algorithm,from the node hardware module failure and node failure rate influence on the accuracy of fault diagnosis of the two aspects in the WSNs node fault diagnosis. The improved ant colony optimization RBF neural network of initial weights is applied to the fault diagnosis of the WSNs node. Parallel optimization characteristics of ant colony algorithm and adaptive adjustment coefficient of volatile characteristics as the clustering algorithm to determine the initial weights of RBF neural network,at the same time adopt crop reduction RBF neural network hidden layer,optimize network structure. Through the experimental results showthat the optimization of RBF neural network based on Ant Colony Clustering Algorithm of WSNs node fault diagnosis method can accurately achieve perception of fault diagnosis,compared with other method has higher diagnosis accuracy.

关 键 词：物联网 节点故障诊断 RBF神经网络 

分 类 号：TH165.3[机械工程—机械制造及自动化] TG65[金属学及工艺—金属切削加工及机床]