Research Of Intelligent Optimization Decision Feedback Blind Equalization Algorithm

Blind equalization algorithm because it did not need the training sequence,and can effectively eliminate the inter-symbol interference(ISI),more and more researchers will serve as a research subject,and put forward many improved blind equalization algorithm.In this paper,based on the analysis of constant modulus algorithm(CMA)and constant modulus decision feedback blind equalization(CMDFE)algorithm,in order to improve the equalization effect,the intelligent optimization algorithm and variable step size theory are applied to CMDFE.The specific research contents are as follows:1.Analyzing constant modulus blind equalization algorithm based on different step function.Based on the theoretical analysis of the constant modulus blind equalization algorithm,the effect of step length on the constant modulus algorithm is analyzed.Aiming at the contradiction between convergence rate and mean square error,this paper analyses the Mean Squared Error based Variable step size Constant Modulus Algorithm(MSE-V-CMA)and Residual Error Autocorrelation based Variable step size Constant Modulus Algorithm(REA-V-CMA).The simulation results show that under this circumstance of guaranteeing convergence,convergence rate and mean square error can achieve the desired effect.2.Analyzing variable step size constant modulus decision feedback blind equalization algorithm.For channel of frequency response with obvious fluctuation,this paper analyses the constant modulus decision feedback blind equalization algorithm(CMDFE),and simulates this algorithm.Because it still exists the contradiction between the convergence rate and mean square error,this paper studies the Mean Squared Error based Variable step size Decision Feedback Equalization algorithm(MSE-V-CMDFE)and Residual Error Autocorrelation based Variable step size Decision Feedback Equalization algorithm(REA-V-CMDFE),and verifies performance of the algorithm by simulation.3.Analyzing the combination of mixed intelligent optimization norm decision feedback blind equalization algorithm.Aiming at the initialization and local convergence problems of CMDFE,on the base of analyzing PSO,SA and AFSA optimization algorithms,we do research two hybrid optimization algorithms:1)SA-PSO-CMDFEBecause the particle swarm optimization algorithm(PSO)is easy to fall into local optimum,and the convergence speed is slow in the late convergence,the simulated annealing algorithm(SA)has a strong ability of global search,it can make up for the disadvantages of PSO in this area.Therefore,the SA algorithm combined with the PSO algorithm apply to CMDFE,a simulated annealing particle swarm optimization constant modulus decision feedback blind equalization algorithm(SA-PSO-CMDFE)is proposed.The superiority of the algorithm is verified by computer simulation.2)SA-AFSA-CMDFEArtificial fish swarm optimization algorithm(AFSA)convergence is slow,and easily trapped in local optima,the combination of SA and AFSA algorithm used in CMDFE,a simulated annealing artificial fish swarm optimization algorithm constant modulus decision feedback blind equalization algorithm(SA-AFSA)is proposed,computer simulation results show that the proposed algorithm has better equalization performance.4.Research on hybrid intelligent optimization variable step Decision Feedback Equalization Algorithm.In order to further improve the convergence of the algorithm,on the basis of previous study,the variable step size theory were applied to simulated annealing particle swarm optimization model of decision feedback blind equalization algorithm based on simulated annealing(SA-PSO-CMDFE)and artificial fish swarm optimization model based decision feedback blind equalization algorithm(SA-AFSA-CMDFE).A variable step size constant modulus decision feedback blind equalization algorithm based on simulated annealing particle swarm optimization is proposed and a variable step size constant modulus decision feedback blind equalization algorithm based on simulated annealing artificial fish swarm optimization(SA-AFSA-VCMDFE).Finally,the validity of the algorithm is verified by simulation experiments.