Blind Separation of Linear-Quadratic Mixtures of Real Sources Using a Recurrent Structure

摘要

In this paper, we propose an approach for separating linear-quadratic mixtures of independent real sources. The method is based on parametric identification of a recurrent separating structure by means of an adaptive algorithm which uses the higher-order statistics of the outputs of this structure. We study the local stability of the recurrent structure and show experimentally that when it is stable at the separating point, it succeeds in separating the sources.