Blind Source Separation Algorithm Based on Wavelet Denoising

In this paper, a method of blind source separation (BSS) is proposed based on wavelet denoising. This method firstly makes wavelet transform (WT) to the observation signals, and then adopts a uniform method which combines signal whitening and maximum information algorithm to obtain the separate signals which approach to the source signals. Simulations show this method could achieve better performance than that without wavelet denoising, and when the iterative number reaches 4000, it could achieve very good results.