Abstract

This thesis focuses on the study of efficient channel estimation techniques for WCDMA communication systems. Some background material pertaining to wireless multipath propagation and WCDMA systems is first presented. The maximum-likelihood principle based data-aided channel estimation techniques are then investigated, resulting in two new estimation methods using pilot and information symbols. The first method combines the pilot-assisted and the decision-directed estimation schemes, yielding a much better estimation result compared to each of the two schemes alone. The second technique corresponds to an adaptive channel estimation scheme in which a new update mechanism using a variable step-size is proposed. A theoretical analysis of the mean square error of the least mean square (LMS) algorithm is conducted, showing the superiority of a variable step-size update scheme in convergence speed over the conventional constant step-size algorithm. The proposed scheme is not only able to increase the convergence speed but also can decrease the steady-state error of the adaptive estimation filter to a certain degree. The bit-error rate (BER) performance of both downlink and uplink of the WCDMA systems using the proposed channel estimation techniques is evaluated through extensive computer simulations for different mobile velocities, confirming the improved estimation accuracy of the proposed methods for both slow to fast fading channels in comparison with some of the existing techniques (Abstract shortened by UMI.)