摘要
提出了基于小波包与量子神经网络的容差模拟电路的软故障诊断,将故障的不确定性数据合理地分配到各类中,减少了故障检测的不确定度,提高了故障检测的诊断率,克服了BP在模糊分类方面的局限性.实验证明,采用QNN与自适应BP神经网络相比,故障诊断率可提高11.89%.
Based on wavelet and muti-level transfer quantum neural network,a new soft fault diagnosis method for analog circuits with tolerance is proposed.It can assign the ambiguity data to the corresponding patterns reasonably,thus reducing the uncertainty of pattern recognition and improving the veracity of pattern recognition.Through experiments using QNN and adaptive BP neural network compared,fault diagnosis can improve the rate of 11.89 percent.
出处
《微电子学与计算机》
CSCD
北大核心
2009年第5期187-189,192,共4页
Microelectronics & Computer
基金
湖南省自然科学基金项目(07JJ6132)
关键词
量子神经网络
量子间隔
小波包
容差
软故障
quantum neural network
quantum interval
wavelet packet
tolerance
soft fault
