Date of Award

Spring 2001

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Electrical Engineering - (Ph.D.)

Department

Electrical and Computer Engineering

First Advisor

Hongya Ge

Second Advisor

Yeheskel Bar-Ness

Third Advisor

Nirwan Ansari

Fourth Advisor

Alexander Haimovich

Fifth Advisor

Jack H. Winters

Abstract

This dissertation focuses on topics of Bayesian-based multiuser detection, space-time (S-T) transceiver design, and S-T channel parameter estimation for direct-sequence code-division multiple-access (DS-CDMA) systems.

Using the Bayesian framework, various linear and simplified nonlinear multiuser detectors are proposed, and their performances are analyzed. The simplified non-linear Bayesian solutions can bridge the performance gap between sub-optimal linear multiuser detectors and the optimum multiuser detector.

To further improve the system capacity and performance, S-T transceiver design approaches with complexity constraint are investigated. Novel S-T receivers of low-complexity that jointly use the temporal code-signature and the spatial signature are proposed. Our solutions, which lead to generalized near-far resistant S-T RAKE receivers, achieve better interference suppression than the existing S-T RAKE receivers.

From transmitter side, we also proposed a transmit diversity (TD) technique in combination with differential detection for the DS-CDMA systems. It is shown that the proposed S-T TD scheme in combination with minimum variance distortionless response transceiver (STTD+MVDR) is near-far resistant and outperforms the conventional STTD and matched filter based (STTD+MF) transceiver scheme.

Obtaining channel state information (CSI) is instrumental to optimum S-T transceiver design in wireless systems. Another major focus of this dissertation is to estimate the S-T channel parameters. We proposed an asymptotic, joint maximum likelihood (ML) method of estimating multipath channel parameters for DS-CDMA systems. An iterative estimator is proposed to further simplify the computation. Analytical and simulation results show that the iterative estimation scheme is near-far resistant for both time delays and DOAs. And it reaches the corresponding CRBs after a few iterations.

Share

COinS