HYBRID DIGITAL SIGNAL-PROCESSING NEURAL-NETWORK CDMA MULTIUSER DETECTION SCHEME

We present a new hybrid digital signal processing-neural network two s tep multiuser detection scheme whose small computational complexity ma kes it attractive for real-time CDMA multiuser detection, An investiga tion on the nature of the local minima of the Optimal Multiuser Detect or's (OMD) objective function leads to the development of an efficient algorithmic stage that can reduce significantly the size of the OMD o ptimization problem. This stage may then be followed by a Hopfield Neu ral Network employed to solve a smaller size residual problem of the s ame form. The performance of the hybrid detector is evaluated via simu lations and it is shown to exceed that of other suboptimal receivers a t a much lower computational cost in both synchronous and asynchronous CDMA transmission cases.