Fast Converging Adaptive MMSE Multiuser Detector for DS-CDMA System

A fast converging adaptive minimum-mean-squared-error (MMSE) multiuser detector is proposed for direct-sequence code-division multiple-access (DS-CDMA) systems with severe near-far problem where the convergence rate of adaptive MMSE detectors for distinct users can be very different. It is shown that by successively cancelling the interference signals of strong power users, the convergence rate of the proposed detectors for weak power users can be significantly increased, which helps to reduce the length of training sequence for tracking. It is also shown that the order of cancellation and several important parameters required for interference cancellation can be determined from the convergence behavior of the proposed detector. Numerical results are presented to show that the proposed detector offers improved performance in various DS-CDMA environments.