Blind Source Separation of a Single Channel Based on Repeated Independent Component Analysis

Independent component analysis (ICA) is a kind of steady algorithm for the separation of blind source (BSS). However, there are three assumptions for the ICA application, one of which is that the number of test channels must be more than that of the signal sources. The limitation brings much inconvenience to practical signal acquisition and processing. In this paper, we propose a new method of repeated independent component analysis (Re-ICA) to realize the separation of blind sources. Under a single test channel, we can increase test channels by means of constructing virtual signal channels and use ICA repeatedly to separate every source signal in turn. Numerical simulation and signal processing of practical data acquisition by one single test-channel show that the proposed method is simple and feasible for operation, and is of great potential in engineering application.