Multi-channel signal separation based on cross-bispectra

Abstract

The authors consider the problem in which they want to separate two (or more) signals that are coupled to each other through an unknown multiple-input-multiple-output linear time invariant system. They prove that the signals can be decoupled, or separated, using only the condition that they are statistically independent, and find even weaker sufficient conditions involving their cross-bispectra. By imposing these conditions on the reconstructed signals, they obtain a criterion for signal separation. A computationally efficient iterative algorithm for solving the proposed criterion, that only involves the iterative solution to a linear least squares problem, is presented.<>