Blind Separations of Nonstationary Sources Based on Spatial Time-Frequency Distributions

Blind source separation (BSS) based on spatial time-frequency distributions (STFDs) provides improved performance over blind source separation methods based on second-order statistics, when dealing with signals that are localizable in the time-frequency (t-f) domain. In this paper, we propose the use of STFD matrices for both whitening and mixing matrix recovery, two stages commonly required in many BSS methods, to provide robust BSS performance to noise. In addition, a simpler method is proposed to select the autoterm and crossterm regions of time-frequency distribution (TFD). To further improve the BSS performance, t-f grouping techniques are introduced and investigated which can be applied to reduce the number of signals under consideration, and even to allow an array to separate more sources than the number of array sensors. With the use of one or more techniques proposed in this paper, significant performance improvement of blind separation of nonstationary signals can be achieved.