Average filtered blind separation of DS-CDMA signals with ICA method

Abstract

In DS-CDMA system, the pseudo random sequences are none correlation and the information sequences of different user's are independent, thus Independent Component Analysis (ICA) is appropriate for the blind separate the DS-CDMA signals with little prior knowledge. Original blind separation method of DS-CDMA signals with ICA utilized only one sample value in each chip. However, in common DS-CDMA communication systems, the sample rate is much faster than the chip rate, so there are many samples during a chip interval. A method of average filtered blind separation of DS-CDMA signals with ICA is proposed in this paper. When noise is stationary, the method proposed could increase the input SNR of ICA greatly. The validity of the method is proved by the simulation results at last.