A neural network based CDMA detector with robust near-far resistance

In code division multiple access (CDMA), high efficiency for spectrum usage and robust jamming resistance have been increasingly important. However the inherent near-far problem is an impediment. Although deliberate power control algorithms can partially solve this problem, the cost is high. In this paper, a compact neural network based detector for CDMA is proposed. By proper array processing and applying hardware annealing technique, the VLSI implementation of optimal multiuser detector (OMD) is suitable.