一种奇异混合信号盲分离的神经网络模型

A Kind of Neural Network Model for Blind Separation of Singularly Mixed Sources

混合信号盲分离问题是一类很难而又具有很强应用背景的问题 ,以往对这类问题的研究均在混合矩阵为非奇异的条件下进行 .本文给出一种神经网络模型及相应算法 。

Blind separation of sources is a problem arising in many practical fields, which is difficult to research. In the existing references, the mixing matrix of sources is always supposed to be nonsingular. This paper presents an algorithm and a corresponding networks to deal with blind separation of sources in the case of singular mixing matrix.