To implement the CDMA multiuser detector by using transiently chaotic neural network

The design of a suboptimal CDMA multiuser detector is a hot topic. We propose to implement the CDMA multiuser detector by using transiently chaotic neural network (TCNN). After going through a transiently chaotic process, the network reaches the global optimum or the neighbor of global optimum. Computer simulation results show that the proposed detector is clearly superior to Hopfield neural network (HNN) based detector and comparable with the optimal detector.