Derivation and optimization of convolutional turbo coded OFDM and DMT modulation schemes

The research work presented in this thesis deals with the study, derivation and optimization of partially con-volutional turbo coded modulation schemes aiming at improving the performance of wireless and wirelinebroadband digital communication systems based on OFDM and DMT technologies. Our efforts are focusedon providing coding methods that yield a good trade-off in terms of performance, complexity, flexibility andbackward-compatibility with the coding methods suggested in the Wi-Fi, WiMAX and DSL standards.We first attempt to improve wireline (DSL) systems by introducing an original multilevel coding scheme,called hierarchical trellis coded modulation (HTCM), based on the hierarchical protection of three non-binarylevels: the first level with a convolutional turbo code (CTC) and the two remaining levels with trellis codedmodulation (TCM). A HTCM scheme can outperform by over 1 dB a complexity-equivalent TCM. However,HTCM schemes do not achieve a significant performance improvement in applications such as DSL that use outerReed-Solomon (RS) coding. As an alternative, we suggest a serially concatenated scheme formed by an outerRS code and an inner bi-level turbo coding modulation (TuCM) that protects the first 2-ary level with a CTCand leaving the second level unprotected. A comprehensive optimization of the TuCM shows that a structureemploying WiMAX’s duo-binary CTC and a codeword spanning roughly 900 subcarriers can achieve 7 dB codinggain for a BER of 10−7. A modification of the latter structure for wireless applications involves not using outerRS coding and protecting the second level of the inner TuCM with a convolutional code (CC). As an example, wepropose a structure combining WiMAX’s CC and CTC, which achieves a better performance/complexity trade-off than current standard Wi-Fi and WiMAX solutions. The design and optimization of our coding structuresled to the development of original tools, such as new theoretical bounds on the error probability of multilevelcoding schemes, and a new algorithm for estimating the free distance of turbo-codes. Finally, we propose aso-called self-protection method to enhance the burst-error correction capability of coded multicarrier systemsoriginally optimized for handling random-errors. The technique efficiently combines various classical concepts,such as SNR margin, erasure-decoding and a novel form of channel interleaving, and can yield significantlyreduced latency as compared to traditional techniques, such as channel interleaving and outer RS coding.