| Bit-iterleaved coded technology has the features of simple-structure, easy-configuration and time diversity. It is not only widely applied in several practical communication systems, but also attractive in the new-generation wireless link transmission as a key technology. But there are still many problems. Especially when Turbo principle is imported, the analysis method is much different from the traditional ones. As for bit-interleaved coded transmission with different channel models, some topics are studied, which include performance bound, labeling design in memory-less channels and its application, simplified structure of the iterative receiver under multi-path channels, the impact of non-ideal channel information and iterative signal-to-noise ratio (SNR) estimation.1. Performance bound is an effective tool to validate the transmission system. Based on the summary of the known ones in Bit-interleaved coded modulation (BICM), the modified expurgated theories are proposed to solve the ambiguity in the foregone one. The methods are put forward to determinate the pair-wise error event and calculate the amendatory Euclidean distance. Their common ground and difference are illustrated and interpreted through representative examples. To simplify the proposed bound, an approximated one is also proposed.2. Harmonic means corresponding to the new bounds are used to study the label of BICM and Bit-interleaved coded modulation with iterative decoding (BICM-ID). A simple tree searching algorithm is proposed to find the Gray label set. Using the new Harmonic means, the optimum Gray labels for high-order QAM are obtained. From the structure point, the labels suited to the iterative decoding are discussed, which has good asymptotic performance. Also the programming of joint Harmonic means without and with error-free feedback are built, which is solved by BSA. The labels are the results of the tradeoff of the convergence performance and the asymptotic. Some new labels are given for 64QAM. As for its application, an example in code-cooperation is listed with its analysis.3. For bit-interleaved coded system in multi-path channel, using extrinsic information transfer (EXIT) chart, Turbo equalization with different equalizers and mappings are analyzed. A strategy of iterative control is proposed. Under the given principle, Turbo equalization can be switched to BICM-ID with some iteration, which reduces complexity of the system.4. The effect of non-ideal channel condition on the iterative receiver is discussed, also SNR estimation. Using EXIT chart, the impacts of unmatched SNR parameter on BICM-ID and Turbo equalization are studied. Using the same method, that of channel state information with error on Turbo equalization is also discussed. Then, the optimization is studied about the iterative receiver with unmatched noise type. BICM-ID is used as an example. A simplified suboptimal algorithm with threshold approximation is proposed. Finally, a class iterative SNR estimation algorithm is proposed using the Expectation Maximization (EM) principle. The estimation method and its simplified version with equal A priori are in it, also that with feedback extrinsic information.The proposed analysis methods and algorithms in this dissertation are validated through theoretical proof and simulation in detail.
|
PDF Full Text Request