The current trend in the cellular industry is toward spectrally efficient cellular systems with ubiquitous service coverage. Efficient power and bandwidth utilization is the design objective of this thesis. The thesis can be divided into three parts. The first part of the thesis focuses on the background, including Chapters 1 to 3, presenting introduction, transmitter-receiver model, channel model and the applied channel coding codes.

The contributions of the thesis constitute the remaining two parts. The first part, including Chapters 4 to 6, focuses on the performance analysis of coded CDMA systems and provides guidelines for coded CDMA systems design. In a coded CDMA system, a natural question is how to allocate bandwidth expansion to channel coding and direct- sequence spreading. Consequently, coding-spreading tradeoff in a DS-CDMA system is first analyzed, we find that the best code rate for the minimum-mean square-error (MMSE) receivers depends on system load and the target bit error rate (BER). When the system load increases, or the target BER becomes more stringent, more bandwidth expansion is suggested to allocate to spreading. By applying different coding codes, we find that the best code rate for MMSE receivers depends on the steepness of the slope of their performance curves. We further claim that there is no coding-spreading tradeoff for the conventional matched-filter (MF) receivers. The obtained results are then applied to Reed-Solomon (RS)/convolutional concatenated coding systems. Some insight of the working mechanism of the concatenated coding schemes is observed and analyzed. It is shown that the inner codes are used to clean the channel initially for the outer codes, which in turn, correct the remain errors to achieve very low decoded BER. Regardless of the wide range of the target BER, the inner codes work at a nearly fixed point. We also consolidate the statement presented in the previous Chapter that there is no coding-spreading tradeoff for the conventional MF receivers. Next, hybrid type-I ARQ is incorporated into the analysis. The best compromise between error detection and error correction capabilities of the RS codes is studied. It is suggested that the decoding sphere radius for error-only decoding, and the effective decoding diameter for error-and-erasure decoding, is 1 or 2 smaller than the maximum possible. value, depending on the target codeword error rate. The obtained results are helpful in designing highly reliable data transmission systems.

The second part of the contributions, including Chapters 7 and 8, focus on adaptive techniques. Adaptive coding is studied first in a slowly fading channel. An enhanced Incremental Redundancy (IR) protocol is proposed to reduce the number of transmis- sions. The improvement over the conventional IR protocol is to adaptively adjust the starting code rate for the initial transmission. Simulation results show that the num- ber of transmissions can be greatly reduced, especially when the signal-to-noise ratio is low. It is also shown that the proposed approaches can closely track the average effective code rate of the conventional IR in a fading environment. In Chapter 8, adap- tive closed-loop power control over Rayleigh fading channel is investigated. MMSE receivers are used instead of the conventional MF receivers. The target power and the filter coefficients are iteratively updated taking into consideration the variations of the multiple access interference. The transmit power is updated using an adaptive stepsize algorithm, where the adapting information is generated based on the power control command history and the estimated mobile speed. Furthermore, a simple approach is proposed for mobile speed estimation considering the statistics of the fading envelope. It is an improvement over the conventional envelope level crossing counter approach.