Performance of hybrid hopfield neural networks with EM algorithms for multiuser detection in ultra-wide-band communication systems

Abstract

An adaptive receiver is proposed for direct-sequence ultra-wideband (DS-UWB) multiple access communication systems to suppress both multiple access interference (MAI) and inter-symbol interference (ISI). Due to high complexity of the optimum multiuser detector (MUD), suboptimal multiuser detectors with less complexity and reasonable performance have received considerable attention. In this contribution, a Hopfield neural network (HNN)-aided MUD is proposed for employment in DS-UWB system, which is capable of achieving a similar performance to that attained by its optimum MUD-aided counterpart at a significantly lower complexity, especially at high user loads. Furthermore, the performance of the proposed HNN-aided system can be further improved if the proposed electromagnetism-like method (EM) assisted HNN (EMHNN) MUD is employed. We demonstrate that the EMHNN aided system is capable of reducing the bit error rate (BER) by up to four orders of magnitude in comparison to the suboptimal MUDs and the MUDs based on the HNN.