Adaptive Filter Algorithm And FPGA Simulation

Adaptive filtering technology is the important component on modern digitalfiltering technology，which can automatically adjust filter parameters on the filteringprocess, and then achieve optimal filtering effect. Adaptive filter algorithm is theimportant algorithm on adaptive filtering technologies, which is design good or badhave a direct impact on the filtering effect with filter. However, the minimum meansquare algorithm (LMS) with its high stability, simple calculation, and achieveeasily, which is widely used in the system engineering, missile technology,aerospace technology, digital communication technology, remote sensingtechnology and other high-end areas.This thesis mainly discussed influence in minimum mean square LMSalgorithm based on the performance of the factors, combined with the existingliterature variable step forward long LMS algorithm, and detailed analysis of thealgorithm of each parameter in the influence of its various performance indicatorsThis paper puts forward three improved LMS algorithm of variable step length, anda genetic algorithm is adopted its algorithm parameter optimization, and adoptedDSPBuilder development tools on its proposed algorithm simulation. It mainlyincludes the following aspects of content.1、This paper analysis the principle of the LMS adaptive filter detailed, andmainly discussed four factors in the influence of LMS algorithm performance index,the algorithm main existences the contradiction between convergence speed and thesteady-state disorders. In order to improve the contradiction of LMS algorithmMany literatures put forward each variable step length LMS algorithm. Its mainsolved algorithm of selecting step length problem, Providing ideas through theliterature, this paper puts forward three improved LMS algorithm of variable steplength, Meanwhile analyzed the influence of index on the parameter of α、 β、mand h. It proposed algorithm simulation through MATLAB, and compared with theexisting algorithm performance index. Simulation results show that the proposedvariable step length LMS algorithm has good convergence and tracking. Improvingthe contradiction between the convergence speed and the steady-state disorders. 2、The performance of variable step length LMS algorithm mainly depends onthe selection of each step factor parameter. For the parameter selection of each stepfactor dependence on the experience of its designers. This paper use geneticalgorithm for the parameters optimization design on the variable step length of LMSalgorithm, and get the optimal value of α、 β、m and h quickly and efficiently,Meanwhile do the computer simulation algorithm with the selection of optimalvalue. The results show that this method has good stability and robustness, Itsalgorithm various performance indicators can meet the requirements.3、The Altera company offered the development tools of DSPBuilder, andmodeled the second improved cosine LMS algorithm variable step length. Finally,Using FPGA simulated the model, The simulation results show that the proposedalgorithm has good filte.