Improved Variable Step-Size and Variable Parameters LMS Adaptive Filtering Algorithm

In order to achieve the best filtering performance and make the filter automatically adjust its coefficients with the changing of its working environment, we need to use adaptive filter technique. To overcome the contradiction of traditional LMS algorithm between convergence speed, convergence accuracy and the selection of step-size, a novel variable step-size & variable parameters adaptive filter algorithm based on modified hyperbolic tangent function was proposed, whose theory and process were worked out and the design principle of adjusting factors were given. Theory and simulation results verified the convergence properties and superiorities of this algorithm under different circumstances. Copyright © 2013 IFSA.