关键词: 混合矩阵, 病态盲分离, 稀疏性, 估计, 搜索重构观测信号

Abstract: Based on the sparse character of the signals，a two-stage method can be used to accomplish Blind Signal Separation（BSS） when the mixing matrix is ill-conditioned case.The key of the method is the estimation of the mixing matrix.A new kind of method of estimating the mixing matrix is put forward.A column of the mixing matrix is estimated by searching observation signals newly constructed and finding some sampling data of certain source dominating.Other columns can be similarly estimated.The method needs less sparseness of signals and less computation of algorithm free of optimizing process.Simulation results illustrate the efficiency and the good performance of the algorithm.The quantitative analysis about the performance of the method is made by many simulation results.

Key words: mixing matrix, ill-conditioned Blind Signal Separation（BSS）, sparseness, estimation, searching newly constructed observation signals