Firstly, a basic reverse TD-SCDMA system is setup, including a transmitter, a receiver and a channel. Within the receiver model, a specific soft Viterbi decoding algorithm is developed. Secondly, the system synchronization requirements and their impact on its performance under realistic conditions are evaluated. The need for uplink synchronization is established and the sensitivity of the system performance to time misalignment is qualified. To mitigate multipath fading, a RAKE receiver and a minimum mean-square-error (MMSE) receiver are considered. Two RAKE receivers are compared and it is observed that the MMSE-based RAKE receiver always outperforms the cross-correlation based RAKE receiver. An MMSE receiver, including fully or fractionally spaced equalizer is also considered. MMSE receivers do improve the performance for some channels, but not in others. Fractionally spaced equalizers outperform the fully spaced equalizer.

Among all the third generation (3G) systems, time-division duplex synchronous code-division multiple access (TD-SCDMA) system is a unique system and has a lot advantages. The performance of the TD-SCDMA system is evaluated using both link level computer simulations and analysis. This dissertation is focused on the performance evaluation and improvement from the receiver model perspective, including Viterbi decoding algorithms, receiver structures, channel estimation algorithms, channel equalization algorithms, and smart antenna techniques.

Finally, the performance of the system with an antenna array consisting of a linear array of four equally spaced omni-directional antenna elements, and a direct matrix inversion algorithm (DMI) using the MMSE criterion, was also evaluated.