Triply iterative algorithm for extracting statistically independent sources from array signals with noises

In this paper, an efficient algorithm is developed for blind separation of multiple independent sources from observed array output signals with noises. First, a novel tri-quadratic cost function is introduced for extracting statistically independent sources from array signals. Then an efficient method is established for solving this cost function. The developed blind source separation algorithm involves three main iteration steps, but each iteration step deals with only a least squares problem. Some numerical simulation examples are provided to demonstrate the effectiveness of the developed algorithm.