Blind multi-User detection in DS-CDMA downlink under Gaussian noise using independent component analysis
Lei Shen,Shiju Li,Fangni Chen,Jun Wei
DOI: https://doi.org/10.1109/ICOSP.2006.345857
2007-01-01
Abstract:For the DS-CDMA downlink, the performance of the receiver based on traditional matched filter will be worsen, due to multi-path interference and multiple access interference (MAI). Applying the independent component analysis (ICA) to blind detection of DS-CDMA signals with multi-path interference, we can get all the source signals without any information of the transmitted users. The basic feature of ICA method is able to estimate a set of independent source signals, but their order is somewhat unpredictable and the desired signal can be obtained directly. Since the spreading sequence of the mobile phone is initially known, this paper proposes a new method to estimate both the source signals and the corresponding spreading codes by ICA. The desired signal can be got by comparing the estimated spreading sequences with the initially known spreading sequence. The performance of the proposed method under Gaussian noise is also analyzed. Illustrative simulation examples are provided at last