A new blind separation approach to ill-condition mixed sources

In this paper, we discuss solution property for blind separation of ill-condition mixed sources and therefore derive a sufficient and necessary condition. Further, we present a new contrast function, through which a new algorithm is proposed to directly separate the sources without time-consuming iterative steps. Under a loose condition, It is proved that any source signals theoretically satisfying the condition can he successfully separated by our method. The computer simulation shows its outstanding performance on blind source separation.

I. Introduction

BLIND source separation (BSS) is a kind of technique recovering the original unknown source signals from the observed mixtures of them under the assumption that the sources are all statistically independent each other. Because this technique needs little prior knowledge of the sources and the mixing channels, BSS has a lot of attractive potential applications in wireless communications, biomedicine, speech signal processing, earthquake reconnoitering, and so forth.