Packet data communications over coded CDMA with hybrid type-II ARQ.

摘要

This dissertation presents in-depth investigation of turbo-coded CDMA systems in packet data communication terminology. It is divided into three parts; (1) CDMA with hybrid FEC/ARQ in deterministic environment, (2) CDMA with hybrid FEC/ARQ in random access environment and (3) an implementation issue on turbo decoding.; As a preliminary, the performance of CDMA with hybrid FEC/ARQ is investigated in deterministic environment. It highlights the practically achievable spectral efficiency of CDMA system with turbo codes and the effect of code rates on the performance of systems with MF and LMMSE receivers, respectively. For given ensemble distance spectra of punctured turbo codes, an improved union bound is used to evaluate the error probability of ML turbo decoder with MF receiver and with LMMSE receiver front-end and, then, the corresponding spectral efficiency is computed as a function of system load.; In the second part, a generalized analytical framework is first provided to analyze hybrid type-II ARQ in random access environment. When applying hybrid type-II ARQ, probability of packet success and packet length is generally different from attempt to attempt. Since the conventional analytical model, customarily employed for ALOHA system with pure or hybrid type-I ARQ, cannot be applied for this case, an expanded analytical model is introduced. It can be regarded as a network of queues and Jackson and Burke's theorems can be applied to simplify the analysis. The second part is further divided into two sub topics, i.e. CDMA slotted ALOHA with hybrid type-II ARQ using packet combining and CDMA unslotted ALOHA with hybrid type-II ARQ using code combining. For code combining, the rate compatible punctured turbo (RCPT) codes are examined.; In the third part, noticing that the decoding delay is crucial to the fast ARQ, a parallel MAP algorithm is proposed to reduce the computational decoding delay of turbo codes. It utilizes the forward and backward variables computed in the previous iteration to provide boundary distributions for each sub-block MAP decoder. It has at least two advantages over the existing parallel scheme; No performance degradation and No additional computation.