Blind source separation method based on diagonalization of correlation matrices and genetic algorithm
Tianrui Li · Systems engineering and electronics · 2005
A new cost function based on the diagonalization of correlation matrices is proposed to measure the independency of output signals. In order to expand the search space and decrease the crosscorrelation among the sub-sources, non-linear transform is made for the cost function. The real coded genetic algorithm is also proposed to search for the optimum solution, which can overcome the drawback of traditional gradient search technique that it is likely to tend to fall into the local minimum. This novel method is applicable to instantaneous or convolutive mixture models with stationary or non-stationary input signals. Simulation results demonstrate that the algorithm not only has fast convergence and high accuracy, but also can improve the output SNR greatly.