Improved variable step size LMS adaptive filtering algorithm and its performance analysis

For existing LMS algorithm could not simultaneously improve the convergence speed and lower steady-state error of contradictions.By building a nonlinear functional relationship between μ(n) and e(n),this paper proposed an improved variable step size LMS adaptive filtering algorithm.Compared with existing algorithms,while the introduction of memory factor λ and control functions values of the parameters β(n),so that the current iteration step were related to the previous step and the former M the square of error.Theoretical analysis and computer simulations show that several common with existing LMS algorithm,the improved convergence rate and steady-state error performance is improved.