Abstract
The multiple access interference (MAI) and the near-far effect are two major problems that limit the performance of the conventional detector in the CDMA communication systems. A lot of power control and multiuser detection schemes have been proposed to overcome these two problems. S. Ulukas and R. D. Yates have investigated the minimum mean square error (MMSE) power control scheme, which is composed of a power control algorithm based on the signal to interference ratio (SIR) measurement and the MMSE interference suppression. They verified that with the MMSE power control algorithm the capacity of active users can be increased as compared with that of the conventional power control algorithm. During the process of calculating the filter coefficients and the power assignment in the MMSE power control scheme, the interference power and the interference covariance matrix must be measured first. In this thesis, we show that the estimation of the interference power and the interference covariance matrix with the detector output information can further improve the performance. Then we study the effect of replacing the MMSE interference suppression with two adaptive schemes of less complexity, least mean square (LMS) algorithm and minimum output energy (MOE) algorithm based interference suppression. We also propose a modified LMS algorithm that incorporates the distinct feature of the MOE algorithm. In such a way, we can improve some disadvantages of the conventional LMS algorithm and hence have better performance.