Detection of DS-CDMA signals based on independent component analysis

In this paper, we propose a new detection scheme, which can be used to remove the inherent permutation and scaling ambiguities of the ICA algorithms used for blind source separation. This has the advantage of enhancing the performance of the ICA algorithms when applied for detection of DS-CDMA signals. ICA based techniques are based on independence of source signals and these conditions are satisfied in communication systems. So, in such applications the techniques based on ICA can be useful. Since the spreading code sequence is the only a priory information available, it is shown that this prior information can be exploited for removing the ambiguity of permutation and scaling in the detection of the signals.