Turbo equalization applied to multiuser detection in a convolutionally coded asynchronous flat rayleigh fading DS—CDMA System

Abstract

AbstractIn this paper we propose a structure of a multiuser receiver which reuses the basic principles of turbo—equalization for ISI canceling. The receiver is mainly made of a centralized MUI (Multi—User Interference) canceler which detects the coded data by means of a Viterbi algorithm. The metrics are computed using Verdu's studies and the decisions are weighted using the log—MAP algorithm. This weighting operation followed by an interleaver block which scatters error—bursts enables the channel decoder to work with analog inputs. The channel decoder uses the log—MAP algorithm to weight the decisions that are fed back, in a closeii loop manner, to the centralized MUI canceler. This scheme is clearly close to turbo—decoding or turbo—equalization principles. Simulation results are presented over both additive white Gaussian noise channel and Rayleigh flat fading channel showing each time that the upper bound represented by the single—user detection scheme case can be approximately reached by such a system.