ITERATIVE TECHNIQUES FOR BLIND SOURCE SEPARATION USING ONLY FOURTH-ORDER CUMULANTS

"Blind source separation" is an array processing problem without a priori information (no array manifold). This model can be identified resorting to 4th-order cumulants only via the concept of 4th-order signal subspace (FOSS) which is defined as a matrix space. This idea leads to a "Blind MUSIC" approach where identification is achieved by looking for the (approximate) intersections between the FOSS and the manifold of 1D projection matrices. Pratical implementations of these ideas are discussed and illustrated with computer simulations.

[1]  Jean-Francois Cardoso,et al.  Eigen-structure of the fourth-order cumulant tensor with application to the blind source separation problem , 1990, International Conference on Acoustics, Speech, and Signal Processing.

[2]  R. Liu,et al.  An extended fourth order blind identification algorithm in spatially correlated noise , 1990, International Conference on Acoustics, Speech, and Signal Processing.

[3]  Jean-Francois Cardoso,et al.  Super-symmetric decomposition of the fourth-order cumulant tensor. Blind identification of more sources than sensors , 1991, [Proceedings] ICASSP 91: 1991 International Conference on Acoustics, Speech, and Signal Processing.