一元及多元信号分解发展历程与展望

Univariate and Multivariate Signal Decomposition:Review and Future Directions

现实世界中,所获得的信号大部分都是非平稳和非线性的,将此类复杂信号分解为多个简单的子信号是重要的信号处理方法.1998年,提出希尔伯特–黄变换(Hilbert-Huang transform,HHT)以来,历经20余年的发展,信号分解已经成为信号处理领域相对独立又具有创新性的重要内容.特别是近10年,多元/多变量/多通道信号分解理论方法方兴未艾,在诸多领域得到了成功应用,但目前尚未见到相关综述报道.为填补这个空缺,从单变量和多变量两个方面系统综述了国内/外学者对主要信号分解方法的研究现状,对这些方法的时频表达性能进行分析和比较,指出这些分解方法的优势和存在的问题.最后,对信号分解研究进行总结和展望.

Most signals obtained in the real world are non-stationary and nonlinear,decomposing such complex sig-nals into several simple sub-signals is an important signal processing method.Since the Hilbert-Huang transform(HHT)was proposed in 1998,after more than 20 years of development,signal decomposition has become a relat-ively independent and innovative important content in the field of signal processing.Especially in the past decade,multivariate signal decomposition methods and theoretical research are in the ascendant,which have been success-fully applied in many fields.However,there is no relevant overview report at present.Therefore,this paper system-atically summarizes the development of signal decomposition theory and methods from both univariate and mul-tivariate aspects.This work analyzes and compares the time-frequency expression performance of these methods,and points out the advantages and issues.Finally,the future research of signal decomposition is prospected and summarized.