Robust Neural Network based Multiuser Detector in MC-CDMA for Multiple Access Interference Mitigation

Abstract

The aim of this paper is to design a multiuser detector technique using artificial Neural Network under multiple access interference mitigation. To improve the performance of Multi Carrier Code Division Multiple Access (MC-CDMA) under multiple access interference, we have used a multilayered perceptron model of Neural Network that has three layers namely, input layer, hidden layer and output layer. The Neural Network is further trained by Levenberg-Marquardt algorithm. This algorithm uses the error function as key factor based on which the weights are adjusted to get the desired output. The Bit Error Rate (BER) performance of the system has been evaluated under Rayleigh and Stanford University Interim (SUI) channels for Binary Phase Shift Keying (BPSK) and Quadrature Amplitude Modulation (QAM) techniques. The proposed Neural Network based receiver is compared with Equal Gain Combining (EGC) and Maximal Ratio Combining (MRC) with varying number of users and Signal to Noise Ratio (SNR). Under SUI channel conditions and for a BER of 10 -3 , the Neural Network based receiver shows an improvement of 1 dB and 8 dB than EGC and MRC receivers, respectively.