An artificial neural network based approach to calculate BER in CDMA for multiuser detection using MEM

Multi user detection is an ongoing research in detection theory. They are used to reduce the effect of Multipath fading. In this paper, MEM (Maximum entropy method has been combined with ANN (Artificial neural network) in order to optimize various parameters of CDMA system. We consider the problem of multipath fading and noise in signals being received in CDMA systems. Among all basic detection schemes being used in CDMA systems, SIC(Successive interference cancellation)is considered as best detector due to its exact BER but with more advanced research it has been observed that some limitations of SIC has not been addressed yet. In this paper, we propose an optimization technique in which final output of MEM is compared with neural network. Final results have been compared with existing MEM respectively. Based on simulations, it has been concluded that Bit error rate has been reduced and accuracy has been achieved up to 85%.