摘要
为保障油料供给的安全性,研究无线传感器网络(WSN)节点故障诊断的可行性策略,提出变精度粗糙集(VPRS)和RBF神经网络结合的故障诊断方法.该方法由运行中的汇聚节点实时获取子节点故障征兆,建立初始决策表,利用VPRS作为前端处理系统,对初始决策表进行约简,删除冗余的、不重要的属性征兆,并将约简后的结果输入RBF神经网络实现节点故障识别.仿真实验结果表明:对于具有显著不确定性的WSN节点故障诊断,该方法能够准确快速地得出诊断结果,鲁棒性和适用性更强.
In order to ensure the security of oil supply,the feasible strategy for fault diagnosis of node in wireless sensor network（WSN） is investigated and a fault diagnosis method based on variable precision rough set（VPRS） and RBF neural network for WSN＇s nodes is proposed in this paper.The procedure of the method is as follows.The sink gets node fault symptoms and forms initial decision table firstly.Then the VPRS theory is used as the front-end processing system to remove redundant and insignificant attribute symptoms for getting a relative minimum condition attribute set,which plays a major role in fault diagnosis.Finally,the relative minimum condition attribute set is input into RBF neural network to identify faults.Simulation results show that the proposed method can accurately and quickly arrive at the decision about the fault diagnosis of node with significant uncertainty in WSN.It also has strong robustness and applicability.
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2010年第7期807-811,共5页
Transactions of Beijing Institute of Technology
基金
北京理工大学基础研究基金资助项目(20070542009)
