Blind extraction of chaotic signals with different variances by using the FastICA algorithm

In blind source separation, it is generally assumed that the sources have unit variance. However, the variances of the sources may be different in practice. The results of using the fast independent component analysis (FastICA) algorithm to realize blind extraction of chaotic signals with different variances are reported in this paper. The correlation coefficient and intersymbol interference (ISI) criterion are used to evaluate the performance, and the impact of the length of a signal frame and the different variances of the chaotic signals to the performance are investigated. It is demonstrated that the correlation coefficient is an effective criterion to evaluate the performance while ISI may be not, and the FastICA algorithm can extract the chaotic signals with different variances effectively.