Independent component analysis, a new concept?
Signal Processing - Special issue on higher order statistics
A blind source separation technique using second-order statistics
IEEE Transactions on Signal Processing
Hi-index | 0.00 |
This paper dealswith the extraction of signals from their instantaneous linearmixtures using time-frequency distributions. Fundamentally, thisproblem is a signal synthesis from the time-frequency (t-f) plane.However with the incorporation of the spatial information providedby a multisensor array, the problem can be posed as special caseof blind source separation. So far, the blind source separationhas been solved using only statistical information availableon the source signals. Herein, we propose to solve the aforementionedproblem using time-frequency signal representations and the spatialarray aperture. The proposed approach relies on the differencein the t-f signatures of the sources to be separated. It is basedon the diagonalization of a combined set of spatial time-frequencydistribution matrices. A numerical example is provided to illustratethe effectiveness of our method.