An Algorithm for Blind Source Separation Based on Infomax and its Stability
Fa Wang · Mathematica Applicata · 2005
In this paper,we propose an independent component analysis (ICA) algorithm——extended infomax algorithm which enables to separate mixtures of super Gaussian and sub Gaussian primary source signals.A stability condition of the proposed algorithm is given and proved.Alternative nonlinear function in the algorithm are derived.The function are self adaptive based on estimating the high order moment of the extracted signals.Compared with the original algorithm,the proposed algorithm has a larger field to use.The effectiveness and performance of the algorithm are verified by the computer simulations.