Detection of multipart DS CDMA signals using neural networks

Abstract

Multi-user interference (MAI) cancellation using hysteretic Hopfield Neural Network (HHNN) receiver for Direct Sequence Code-Division Multiple Access (DS-CDMA) in multipath fading channels is investigated. We have shown that by applying the phenomenon of "hysteresis" to the HNN detector, we may enhance the performance of this detector in all near-far situations, for different number of multipath rays. The introduction of hysteresis concept into HNN has made the CDMA HNN detector (abbreviated as HHNN) even closer to the CDMA optimum multiuser detector. As shown by simulation results, the BER performance achieved by the HHNN detector outperforms the classical HNN detector with a good margin and is promising. © 2006 IEEE.