Research On Multi-user Detection In MIMO-OFDM System Based On Ant Colony And Particle Swarm Optimization Algorithm

The rapid development of modern mobile communication technology accelerate theprocess of the communications industry into the B3G/4G times. New generation mobilecommunication systems surpass the original generations because of its outstanding technicaladvantages. The core technology in4G——multi-input multi-output (MIMO) technology andorthogonal frequency division multiplexing (OFDM) technology have also become a hot topicin the field of communication research. A combination of both can be effective against theeffects of multipath and frequency selective fading, significantly save the spectrum resourcesand increase system capacity.In multi-user MIMO-OFDM communication system, the phenomenon of multiple accessinterference is more prominent due to the increase in the number of users. Multi-user detectiontechniques are commonly used to solve this problem. The optimal multi-user detectionalgorithm is maximum likelihood (ML) algorithm which has the excellent performance, but itscomplexity is too high to be applied in practice. Traditional minimum mean square error(MMSE) detection algorithm has the poor performance, and the problem of high error ratealways exist. Therefore, in this paper, ant colony optimization algorithm and particle swarmoptimization algorithm are applied in the detection of multi-user MIMO-OFDM systems inorder to guarantee higher detection performance in low complexity conditions.Ant colony optimization (ACO) is suitable for handling the discrete problem. The positivefeedback mechanism of the algorithm can guarantee that the ant colony find out the optimalsolution. Particle swarm optimization (PSO) has simple concept, fast convergence and a widerange of applications. The algorithm can take advantage of local information and globalinformation to search for solutions. Swarm intelligence algorithm is easily affected by theparameters. So firstly, a large number of simulation experiments are implemengted to select thealgorithm parameters. By the method of controlling variables, we were obtained the optimalparameter configuration of the two algorithms in multi-user detection problem. Then thesimulation results of various algorithms show that ant colony algorithm and particle swarm algorithm can get higher detection performance than that of MMSE algorithm with a certaindegree of complexity of the cost, but still be not close to the performance of ML detection.In order to further improve the detection performance, according to the commoncharacteristics of the groups of ant colony and particle swarm optimization, this paper set theproblem of multi-user detection in MIMO-OFDM system, proposed a new algorithm AC-PSOwith ant colony and particle swarm optimization hybrid. It has better global searching abilityand faster convergence rate, and can effectively avoid premature stagnation. The simulationresults show that the multi-user detection in MIMO-OFDM system based on ant colony andparticle swarm hybrid algorithm can get close to the detection performance of the MLalgorithm and has advantages of the computational complexity beyond the PSO algorithm.Finally, this paper derived the multiplication and addition computational complexity ofACO, PSO, and AC-the PSO algorithm, quantitative analysised the convergence of the hybridalgorithm, and from a mathematical point further proof of the complexity andconvergenceadvantages of AC-PSO hybrid algorithm.In summary, in the problem of multi-user detectionin MIMO-OFDM system, this paperpresented ant colony particle swarm optimization AC-PSO to ensure that the algorithm getclose to the multi-user detection performance of the optimal detection ML algorithm with theaccepted complexity. And this algorithm is an effective algorithm which compromisecomputational complexity and detection performance.