Blind multiuser detection with space-time adaptive processing for CDMA wireless communications.

摘要

Multiple access by code division multiple access (CDMA) has been widely adopted by the emerging wireless systems worldwide. Two major limiting factors of CDMA wireless communications are Multiple Access Interference (MAI) and Inter-Symbol Interference (ISI). To meet the performance and capacity requirements of next-generation digital wireless systems, it is widely considered essential to utilize space-time processing with arrays of antennas (smart antennas) so as to exploit the spatial diversity of wireless systems.; This research deals with the combined space-time signal processing at the receiver for CDMA wireless communications. In particular, it focuses on linear receiver design for multiuser systems using blind detection techniques. The focus is on using appropriate constraints on the detector weights to improve the detection performance in terms of MAI reduction and ISI equalization. Preserving the desired signal energy and suppressing interference are the key concepts in constraint design. Essential constraints include a set of linear constraints for interference suppression and a quadratic inequality constraint for system robustness.; Based on different designs of constraints, we formulate various blind detection methods and derive their adaptive implementations under the frameworks of the minimum output energy (MOE) detection and the decision directed minimum mean squares error (MMSE) detection. These formulations are presented for different system setups, including the single-channel case and the space-time multi-channel case. Under the multiple-antenna setup, we construct both joint and separate space-time receivers, each offering three levels of optimality by using different constraint parameters. The proposed generic constraint design framework provides a comprehensive treatment of the blind MMSE detection mechanism, and facilitates receiver design based on performance and complexity tradeoffs.; The second part of my research is the development of efficient adaptive solutions to the constrained optimization problems that merge in the considered detection formulations. A number of new adaptive algorithms are proposed, including a variable loading technique for implementing a quadratic constraint in quadratic programming and recursive least squares solutions to the max/min optimization problem. These adaptive methods apply both to the detectors that we have discussed, and to the general field of robust adaptive constrained filtering.