A New Variable Step Size LMS Adaptive Filtering Algorithm

Focus on the inconsistency between the convergence rate and the steady state error in LMS algorithm, this paper proposed a new variable step size LMS(VSS-LMS) algorithm, by constructing a nonlinear function between the step factor µ and the error signal e(n). The algorithm which modified the weight vector appropriately achieving the optimal by adjusting the step size parameters, could improve the convergence rate and steady-state error performance effectively. The simulation analysis showed that, this algorithm had faster convergence rate speed and smaller steady-state maladjustment compared with the basic LMS algorithm and parts of the similar variable step size algorithm, it fatherly validated this algorithm was superior to existing algorithms.