Sample-by-sample adaptive space-time processing for multiuser detection in multipath CDMA systems

Multiuser detection and space-time processing are two advanced signal-processing techniques for the mitigation of multiple-access interference and intersymbol interference in wireless CDMA communications. The purpose of this work is to investigate techniques for efficient space-time multiuser detection (ST MUD). A companion paper considered batch iterative methods, which assume knowledge of all signals and channels. In this paper sample-by-sample adaptive methods, both data-aided (with training sequences) and blind, which require only the timing and training sequences (for data-aided) or the spreading codes (for blind) of the desired user(s), are considered. For data aided adaptive methods, a decentralized adaptive minimum-mean-square-error space-time multiuser detector and a centralized adaptive decision-feedback space-time multiuser detector are presented. Then a blind adaptive space-time multiuser receiver based on the linear constrained minimum variance criterion and min-max parameter estimation is developed, which is robustified with norm-constrained techniques in the case of signature waveform mismatch. A least mean square implementation of all these adaptive ST MUD receivers is given.