摘要
针对电子系统,尤其是模拟电路的早期故障诊断是重要却又困难的问题,提出一种基于隐马尔可夫模型(HMM)的模拟电路早期故障诊断方法.首先提取出模拟电路的电压特征;然后用改进的线性辨别分析法(LDA)对电压特征进行降维并消除其冗余,采取一些改进措施来消除LDA的不足;最后将改进LDA提取的特征构成多个观测序列并用于训练和测试HMM,以实现模拟电路的早期故障诊断.将该方法与其他方法进行比较的实验结果表明,其具有良好的故障识别能力.
Diagnosis of incipient faults for electronic systems especially for analog circuits, is important yet difficult. The paper proposes a new method using HMM for the diagnosis of the incipient faults in analog circuits. First, the voltage feature vectors are extracted from analog circuits; Then, the improved LDA is used to reduce dimensionality of the voltage feature vectors and remove their redundancy, where the performance of LDA is improved through overcoming shortcomings existing in original LDA; Finally, the processed feature vectors are used to form the observation sequences, which are sent to HMM to accomplish the diagnosis of the incipient faults. The experimental results on performance of the proposed method indicated that the method has better recognition capability than the other methods.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2010年第7期1215-1222,共8页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(60673011)
国防基础预研项目(A1420061264)
四川省教育厅重点项目(08ZA067)
关键词
隐马尔可夫模型
线性辨别分析
特征提取
故障诊断
hidden Markov model
linear discriminant analysis
feature extraction
fault diagnosis
