Adaptive Blind Separation Algorithm Based on Covariance Matrix
Chuang-Xin Chen · Acta Scientiarum Naturalium Universitatis Sunyatseni · 2005
Propose a new blind separation algorithm is proposed for linear mixture of signals.The algorithm's separation criterion is based on the diagonal characteristic of covariance matrix of the mutually independent signals.Separation is achieved by steepest descend method.The algorithm does not needs extra demand of signals and mixing matrix.Theoretical analysis and simulations show the effectiveness of this algorithm.