Multiuser Detection For Wideband Wireless Communication Systems

Wideband wireless communications is one of the most vibrant areas in the communication field today. Because of the merits of MIMO-OFDM and UWB techniques, they have been comprehensively considered as the major candidates of the future wideband wireless communication system. It is effective that the multiuser MIMO-OFDM and UWB systems allow one single base station communicating with many users simultaneously. However, the fading in wireless channel leads to interference between users. To deal with this problem, multiuser receiver could be employed, namely multiuser detection technique. Thus, this dissertation investigates multiuser detection technique in the two kinds of future wideband wireless communication system above. The main contents of this dissertation are outlined as follows.1. This dissertation exploits the decision feedback detection, and proposes a kind of non-linear multiuser detection algorithm based on decision feedback equalizer (DFE) for MIMO-OFDM system. First, the dissertation analyzed the multiple feedback structure of the proposed DFE-based algorithm. Then, it compares the minimize mean square error (MMSE) performance of the proposed DFE-based algorithm with the traditional DFE. Calculation results proof that our proposed DFE-based algorithm is better than the traditional DFE in the sense of MMSE performance. Finally, according to the known coefficients of previous matched filter, the dissertation puts forward a new coefficient optimization algorithm (COA) for the equalizer. Simulation results show the DFE-based algorithm has the better bit error rate (BER) performance and larger system capacity than the traditional multiuser detection algorithms.2. Indoor UWB channel for short-range wireless communication system has been measured, and its multipath fading-model can be obtained. In this dissertation, the lognormal fading multiuser UWB system model has been established, and a novel genetic algorithm based on complementary error function mutation (CEFM-GA) has been proposed for multiuser detection in the lognormal fading UWB channel. The dissertation discusses the procedure of the CEFM-GA, and analyzes the mutation operation, namely using complementary error function mutation to calculate the mutation probability for 1-dimensional (1-D) scenario. Simulation results show the CEFM-GA can obtain better BER performance under the IEEE 802.15.3a channel.3. This dissertation also establishes the Nakagami fading multiuser UWB system model. Weighting the advantages and disadvantages between the genetic algorithm (GA) and the ant colony optimization algorithm, the dissertation proposes a hybrid algorithm of complementary advantage for multiuser detection, i.e. ant colony optimization combining genetic algorithm based on Q-function mutation (ACO QFM-GA). The ACO QFM-GA translates the results of the GA into the pheromone distribution of the ant colony optimization algorithm in the early period; then employs the feedback of the ant colony optimization algorithm to obtain the final results. The procedure of the proposed the ACO QFM-GA can be divided into two parts. One is the GA which is the first stage of the ACO QFM-GA. In this part, the dissertation detailedly analyzes the mutation operation of the GA which is based on Q-function mutation. The other part is ant colony optimization algorithm which is the second stage of the ACO QFM-GA. In this part, the dissertation concretely investigates the ant colony optimization algorithm. Simulation results show the ACO QFM-GA can attain relative low BER under the condition of relative low computation complexity in the Nakagami fading line-of-sight channel.4. According to the characteristic of iteration calculation, this dissertation proposes a kind of reduced computation complexity approach, i.e. differential algorithm (or differential multiuser detection algorithm). In the multiuser MIMO-OFDM system, the dissertation proposes a group of DFE-based differential multiuser detection algorithms. Firstly, basic idea of differential algorithm has been interpreted and the differential structures of the proposed algorithms have been described. Secondly, the differential algorithms for different kinds of modulation have been analyzed. Finally, the computation complexity of different multiuser detection algorithms can be obtained by float-point operation method. Simulation results show the proposed differential algorithms can save a lot of computations compared with their former algorithms without differential structures. Also, the differential algorithm is the linear transformation, so it hardly affects BER of the detection system.5. The GA has a fast convergence towards the optimal solution, though its computation complexity limits the performance of GA-based multiuser detector. Thus, this dissertation takes CEFM-GA as an example, and proposes a GA-based differential multiuser detection algorithm (i.e. CEFM-GA DA) for multiuser UWB system. Firstly, the differential structure of the CEFM-GA DA has been described. Then, the principle of the differential algorithm in CEFM-GA DA has been thoroughly investigated. Finally, the computation complexity of CEFM-GA DA and other multiuser detection algorithms have been calculated by float-point operation method. Comparing to CEFM-GA, CEFM-GA DA not only has a much lower computation complexity, but also it can indirectly achieve better BER performance by differential algorithm. Incidentally, besides the research on multiuser detection, the proposed DFE-based detection algorithm could make a positive impact on pervasive study of DFE. Also, theoretically, the differential algorithm could be equally applicable to solve the computation complexity for the system which has the characteristic of iteration calculation.