基于小波降噪和盲源分离算法的信号分离方法研究

摘要

Aiming at the problem of blind source separation in noisy environment,wavelet de-noising method is introduced into blind source separation model.Two separation methods,pre-denoising+blind source separation algorithm,and pre-denoising+blind source separation algorithm+post-denoising,are analyzed theoretically.Then,comparison is made to the separation results of the two methods through digital simulation.The simulation shows that:1) Compared with the method only using blind source separation algorithm,the separation method combining wavelet de-noising with blind source separation algorithm has much better separation effect;and 2) The separation effect of pre-denoising+blind source separation algorithm+post-denoising is generally better than that of pre-denoising+blind source separation algorithm.%针对含噪情况下的盲源分离问题,将小波降噪方法引入盲源分离模型之中,对两种小波降噪与盲源分离算法结合的信号分离方法,即预降噪+盲源分离,以及预降噪+盲源分离+后降噪,从理论上进行了原理与特点分析,并通过数值仿真比较了不同信噪比情况下两种方法对于混合信号的分离效果.仿真表明,采用小波降噪与盲源分离结合的信号分离方法,相比于单纯采用盲源分离算法,能够大大提高信号的分离效果;预降噪+盲源分离+后降噪的信号分离方法比采用预降噪+盲源分离的信号分离方法效果普遍提高.