An Improved LMS Adaptive Filtering Algorithm and Its Analysis

The advantage of the traditional fixed-step LMS algorithm is that it is simple and easy to implement, but its inherent disadvantage is that it increases the step factor to speed up the convergence and also causes a large steady-state error. Based on the analysis of the performance of LMS algorithm, an improved LMS algorithm based on the nonlinear relationship function between μ(k)and x(k) is proposed. The theoretical analysis and simulation results show that the vμ LMS algorithm can be compared. Good to overcome the inherent flaws of the LMS algorithm.