基于神经网络的混合电路故障检测  

Fault Detection of Mixed-Circuit Based on Wavelet Neural Network

作　　者：王恩亮[1] 沈元隆[1] 

机构地区：[1]南京邮电大学电子科学与工程学院

出　　处：《通信技术》2010年第6期241-242,245,共3页Communications Technology

摘　　要：随着电路越来越走向集成化和模块化,大多数电路都同时包含数字和模拟的混合信号,数模混合电路故障诊断技术也因此成为一个重要的课题。提出了利用小波分析和神经网络对混合集成电路故障进行检测的方法。基本思想是通过小波变换对原始采样信号进行检测,再利用神经网络对小波变换的结果进行分类,最后给出故障的信息。通过MATLAB仿真实验,证明该方法对混合电路的故障检测非常有效。With the gradual integration and modularization of the circuits,most of the circuits contain both digital and analog signal,and the fault diagnosis technology of mixed digital-and-analog circuits has become an important issue.This paper proposes a fault detection of the mixed circuits based on wavelet analysis and neural networks.The basic idea is that,through the method of wavelet transform,the original sample signal is tested,and then by the neural network,the results from the wavelet transform are classified,and finally,the fault information is given.The simulation experiments with MATLAB show that the method is quite effective for the mixed-circuit fault detection.

关 键 词：小波分析 神经网络 小波神经网络 混合电路 

分 类 号：TP306.3[自动化与计算机技术—计算机系统结构]