Blind separation of underwater acoustic signals via independent component analysis algorithms

Independent component analysis (ICA) algorithms embody an emerging method for blind separation of signals. Present paper applies several ICA algorithms to separate underwater acoustic signals. Simulating with artifical signals, five ICA algorithms are analyzed and compared. The shortage of these algorithms to be used to separate underwater acoustic signals is also pointed out. Based on a noisy ICA model, a method is proposed which separates underwater acoustic signals in presence of noise successfully.