Research On The Multi-user Detection Technology In The MIMO-OFDM System Based On The Hybrid Genetic Algorithm

With the rapid development of mobile communications, the orthogonal frequency division multiplexing (OFDM) technology and the multiple input multiple output (MIMO) technology have become the research focus in the communications area. The OFDM technology and the MIMO technology have complementary advantages, and they can combine into the MIMO-OFDM technology, which is the first choice scheme of the next generation mobile communication system. In addition, the MIMO-OFDM technology can restrain the multi-path fading and the frequency selective fading effectively, also can improve the data transmission rate, the frequency Spectrum utilization rate, and the system capacity.In the mobile communication systems based on the MIMO-OFDM technology, as the number of users and the communication traffic grow larger, the multi-user detection (MUD) technology has aroused the attention of many researchers. Based on the principles of the MIMO-OFDM technology and the MUD technology, the performance of two conventional MUD algorithms, the minimum mean square error (MMSE) algorithm and the maximum likelihood (ML) algorithm, are simulated respectively. The simulation results show that the performance of the MMSE-MUD algorithms is much lower than the ML-MUD algorithms. Also, the simulation results prove that the traditional MUD algorithms have some limitations, such as, the performance of the MMSE algorithm is lower, and the complexity of the ML algorithm is higher.Aiming at the traditional MUD algorithm cannot consider the performance and the complexity simultaneously in the MIMO-OFDM systems, the genetic algorithm (GA) which is based on the MMSE algorithm and the ML algorithm is proposed to solve this problem. The GA algorithm uses the results of the MMSE as the initial population, and uses the criterion of the ML as the fitness function, then generates new genetic population by means of some genetic operators, such as the roulette wheel selection operator, two points crossover operator, adjacent bit reverse mutation operator. The simulation results show that the GA-MUD algorithm in the MIMO-OFDM systems can improve the performance of the MMSE-MUD algorithm under the premise of increasing the complexity appropriately, in addition, the performance and complexity of the GA-MUD algorithm are still lower than the ML-MUD algorithm.The GA algorithm can combine with other intelligent optimization algorithms to be hybrid genetic algorithm (HGA), for example, the GA algorithm can combine with the simulated annealing (SA) algorithm or the particle swarm optimization (PSO) algorithm to be the genetic simulated annealing algorithm (GSAA) or the genetic particle swarm algorithm (GPSA), where the SA algorithm has stronger local search ability and the PSO algorithm has stronger population search ability. The GSAA-MUD algorithm and the GPSA-MUD algorithm are proposed to solve the MUD problem in the MIMO-OFDM systems respectively. The simulation results show that these two HGA-MUD algorithms in the MIMO-OFDM systems can further improve the detection performance, while reduce the complexity of the GA-MUD algorithm.Taking above research on the HGA-MUD algorithm in the MIMO-OFDM system into consideration, the extended study of the HGA algorithm is done from the view of mathematics. First, take the GSAA algorithm for an example to demonstrate that the HGA algorithm follow its schema theorem which is similar to the GA algorithm. Then, the complexity and the convergence of the HGA algorithm and the correlation of the fitness function are analyzed quantitatively. The correlative simulations in the MIMO-OFDM system can prove the correctness of above analysis, and provide a theoretical support for the application of the HGA algorithm.In summary, the HGA-MUD algorithm in the MIMO-OFDM system can gain much higher performance than the MMSE algorithm, and approach to the ML algorithm at the cost of appropriate complexity. In other words, the HGA-MUD algorithm can better balance the complexity and the performance in the MIMO-OFDM system.