FastICA-Based Blind Separation of Digital Modulation signals for Radio Surveillance

In this paper, we propose a blind signal separation procedure for digital modulation signal using FastICA algorithm. FastICA is one of ICA (Independent Component Analysis) algorithms and it enables very fast optimal weight calculation for signal separation without any the prior information of reference signal. ICA is based on the independence of signal to each other. ICA simultaneously separates all the observation signals by array sensors (antennas, microphones, etc) if the number of incident signals is less than the number of elements. In this paper, we first analyze the undesired phase rotation of the separated signal by FastICA, that happens for complex input signal, and study the compensation technique for the rotation. Moreover, we evaluate the separation accuracy and the computational cost of the proposed algorithm in comparison with some conventional algorithms through computer simulation.