Improved blind source separation method based on independent component analysis and empirical mode decomposition

Blind source separation (BSS) has recently received a great deal of attention in signal processing. In order to improve the limited separation performance of the conventional BSS method by the influence of the probability density of the source signal, based on the Independent Component Analysis and Empirical Mode Decomposition theories, an improved blind separation method is proposed. The method is demonstrated by some examples. Simulation results show the improved separation performance of the proposed method, and the time-frequency feature of the source signal has a better reflection.