Research and simulation of blind signal separation technology

The blind signal separation (BSS) is a powerful signal processing method in the biomedical signal processing, array signal processing, speech recognition, image processing, mobile communications and other fields, which obtaining a wide range of applications. The natural gradient algorithm is an important method among the BSS algorithm. This paper mainly introduces the analysis of natural gradient algorithm and realizes application in the speech signal processing. Through simulation, the natural gradient algorithm gets a good signal separation result, which separating the initial signal from the mixing blind signal.