Research on fault diagnosis in wireless sensor network based on improved wavelet neural network

被引：0

作者：

Li, Jie ^{[1
]}

Chen, Bin ^{[2
]}

机构：

[1] Shanghai University of Engineering Science, Shanghai,200437, China

[2] Wenzhou Power Supply Company, Wenzhou,325000, China

来源：

Acta Technica CSAV (Ceskoslovensk Akademie Ved) | 2016年 / 61卷 / 02期

关键词：

Failure analysis - Fault detection;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

Application of wavelet neural network in fault diagnosis of WSN is studied. As classical wavelet neural network algorithm adopts gradient algorithm, it usually has low convergence rate and easily falls into local minimum. To solve this problem, an improved wavelet neural network based on additional momentum and adaptively-Adjusted learning rates is proposed. The results of training experiments show that the improved algorithm has faster convergence speed. Finally, the feasibility and good fault-Tolerant performance of the improved algorithm in fault diagnosis of WSN are verified by simulation experiments. © 2016 Institute of Thermomechanics CAS, v.v.i.

引用

页码：117 / 129