Study On The Technology Of Blind Separation Of Communication Signals

Modern communication environment has become increasingly sophisticated,especially in the field of communication detection in which received signals are oftenmixed with a variety of interfering signals. As the priori information of the sourcesignals and the transmission channel is unknown, the traditional signal processingmethods have shown obvious shortcomings. Blind source separation (BSS) has solvedthis problem well and brought new ideas and methods to separate communicationsignals blindly.Based on theoretic introductions of BSS, this paper detailedly analyzes two kindsof traditional adaptive blind source separation algorithms: the FastICA algorithm andthe natural gradient algorithm. By means of the simulation and after a thorough analysisof the two algorithms, it has proved the FastICA algorithm has the faster convergence,the smaller amount of calculation and the higher efficiency.Due to the issue that traditional separation algorithms couldn’t work effectively inintercepting enemy signals in the field of practical communication detection, this paperfocuses on discussing the technology of BSS in the single-channel based on theunderdetermined blind source separation. Meanwhile a single-channel blind sourceseparation algorithm based on maximum a posteriori estimation is also improved. Thealgorithm could firstly construct the source signals into a series of basis functionsthrough the ICA theory and then in the phase of signal separation estimate the sourcesignals through maximum posteriori theory with the basis functions as prioriinformation. Simulation results have proved the efficiency of the improved algorithm inseparating single-channel signals.