Algorithms research on nonlinear blind source separation
He Li · Sciencepaper Online · 2009
The origin,development and early results of nonlinear blind source separation(BSS)research are mainly described in this paper.The two main types of research methods on non-linear blind source separation are introduced:one is directly extraction non-linear component method based on self-organizing map and another is the method based on assumable mixed-model,the three assumable mixed-models are analyzed in details in the second method.Post non-linear mixing model is extensively studied and is solvable model.The research results based on this model method are commented in recent 20 years.Finally,the existing problem and development tendency about nonlinear BSS are overviewed and given.