基于联合时频分析的混合神经系统在信号分类与模式识别中的应用

介绍了一种基于STFT,WVD和Wavelet变换的混合神经系统在特征提取和信号分类中的应用.它运用时频信号处理技术与混合神经网络结构以及高级训练算法相结合的设计思想,辅之以主元分析和全局决策融合策略,对传统的模式识别技术进行了有效地改进.本文将所述基于WT,STFT和WVD的混合神经系统信号分类器(WSWHNS)的算法程序嵌入一汽车实时智能故障诊断软件包中做了现场实验,获得了非常满意的诊断效果.

Application of a Hybrid Neurosystem Based on Joint Time-Frequency Analysis for Signal Classification and Pattern Recognition

This paper introduces the application of a joint time frequency analysis based hybrid neurosystem trained to provide efficient feature extraction and signal classification. The system, which uniquely employs the combination of wavelet transform, short time Fourier transform and Wigner Ville distribution, hybrid neural network architecture and advanced training algorithms, principal component analysis and global decision fusion in its design, provides significant improvement over traditional pattern recognition techniques. We present our theories' practical application on modern automobile real time intelligent fault diagnosis, and the diagnosis results verify the correctness and effectiveness of our theories, algorithms and programs.