Performance analysis of space-time multiuser detection with error correction coding in CDMA communication systems

With the ever growing demand in digital wireless communications, methods have to be found to improve the performance of such systems. In direct-sequence code-division multiple-access (DS-CDMA), the minimum-mean square-error (MMSE) multiuser detector (MUD) has shown most of the attention in this quest due to its joint capability of suppressing multiuser interference and combating additive Gaussian noise. In addition to multiuser detection, adaptive antenna processing can play an important role in improving the system performance. Furthermore, the use of channel coding, as an essential part of any communication system, in space-time multiuser detection has been lacking in-depth analyses. Issues such as capacity gain improvement due to channel coding, effect of channel fading on the performance, and channel interleaving in DS-CDMA space-time MUD have not been considered in the literature. In this thesis, we attempt to fill this void by providing an accurate and rigorous analysis of such systems.; First we consider a synchronous DS-CDMA system with space-time MMSE-MUD and two types of practically available forward-error-correction (FEC) codes, Reed-Solomon (RS), and convolutional codes (CC). We obtain the BER using the Gaussian approximation for the uncoded space-time system. Then, we develop error bounds for both coding schemes. Using both system simulation and error bounds, we prove the superiority of soft-decision (SD)-CC over RS at all practical bit-error rates (BER's).; Second, we consider both synchronous flat fading, and asynchronous frequency selective-fading channels in a DS-CDMA system with space-time MMSE-MUD and channel coding. In our analysis we obtain the uncoded BER and error upper bounds for the coded system. These analytical results are shown to be in a good agreement when compared to system simulations. The usefulness of our analytical results is made clear in performing capacity investigations where a large number of users is considered and the use of system simulations is out of the question. For instance, we show that capacity gain improvements of approximately 250% and 300% can be achieved using SD-CC for the flat and the frequency selective-fading channels, respectively. We also show that using a practical interleaver can get us close to the ideal interleaving case.; Also we consider a simple adaptive implementation for the MMSE-MUD receiver where the learning period of the adaptive filter is significantly shortened using iterative MMSE and parallel interference cancellation (PIC) algorithms. On a flat fast-fading channel, we show that the BER analysis developed for the multiuser-type receiver can serve as an optimum performance measure when steady state is achieved in the adaptive receiver. Based on these findings, we show that using a recursive-least-squares (RLS) algorithm with a suitable channel interleaver, a performance close to the multiuser-type receiver can be achieved with a capacity gain of approximately 225% relative to the uncoded space-time adaptive receiver. Finally, in Appendices D and E we study the use of turbo codes in space-time channels using only system simulations.