Blind Signal Separation Using an Extended Infomax Algorithm

The Infomax algorithm is a popular method in blind source separation problem.In this article an extension of the Infomax algorithm is proposed that is able toseparate mixed signals with any sub- or super-Gaussian distributions. This ability isthe results of using two different nonlinear functions and new coefficients in thelearning rule. In this paper we show how we can use the distribution of observationvectors for selecting suitable coefficients for our algorithm. Hence, the proposedalgorithm is suitable for real applications in which the distribution of source signalsmight be unknown. It is also shown in this paper that the extended Infomaxalgorithm is able to separate 23 sources with a variety of distributions. Incidentally,we use a performance criterion for the evaluation of our results, based on thecomparison of Kurtosis of the original signals and estimated signals.