Research On Turbo Codes And Multiuser Detection In Wideband Mobile Communications

As is well know that Turbo codes and Multiuser detection (MUD) are among the most important techniques in the third generation (3G) mobile communication, Turbo codes are recognized as having the performance closed to the Shannon limit and be regarded as the milestone on channel coding theory. Turbo codes make full use of the modern coding principles such as concatenated coding, soft input soft output decoding and iteration decoding, in which iteration approach has been applied in many fields like channel equalization, detection of interference between symbols as well as multiuser detection. Multiuser detection (MUD), also a hot topic in wideband mobile communication, is mainly concerned with increasing the numbers of the user and enhancing the quality of the whole communication system.In the thesis, from a new point of view, we investigate the application of non-linear dynamics into the Turbo codes and MUD area, namely, how the chaos dynamics can be applied in Turbo codes development, and how the chaos and neural networks dynamics can be applied in MUD.The main contribution of the thesis can be summarized as follows:1. An iterative decoding algorithm based on MAP (maximum a priority) and SOYA (soft output of Viterbi algorithm) in AWGN (additive white gauss noise) channel is proposed, and the effectiveness of the algorithm is justified by simulation performed on AWGN and Rayleigh fading channels.2. Three criteria for evaluating performance of interleaver, ie., the codeword weight decision, correlation analysis and iterative decoding suitability (IDS), is proposed. The chaotic interleaver is proposed based on chaotic dynamics. It's proven that novel the chaotic interleaver has favorable performance and other advantages verified by computer experiment simulations.3. A new Turbo codes using chaotic soft spreading spectrum is also suggested. By theoretic analysis and simulation, the novel codes has better performance such as lower BER (bit error rate) and other features as compared to conventional Turbo codes. On this basis, a new codes called RSI (randomspreading iterative) codes with the capability of approaching the Shannon limit is introduced.4. By mapping MUD to a stable feedback neural network, a novel MUD approach in AWGN channel is proposed. The theoretical proof and computer simulations justify the validity of the method. Referred to the real fading channel, a novel MUD technique based on mean field theory is presented. The theoretical proof and the simulation verify the effectiveness of the method loo. Referred to the amplitude fading and phases delay in complex channel in reality, a new complex neural network MUD method is suggested for flat Rayleigh fading channel. Besides, a chaotic MUD method is also presented. The two MUD approaches have been justified by theoretical analysis and computer simulations.