An improverd variable step size LMS adaptive filtering algorithm

LMS (Least Mean Square) algorithm is widely used due to its simple and stable performance. As is well known, there is an inherent conflict between the convergence rate and stead-state misadjustment, which can be overcome through the adjustment of size factor. The paper has analyzed some LMS algorithms that already existed and a new improved variable step-size LMS algorithm is presented. The computer simulation results are consistent with the theoretic analysis, ?which show that the algorithm not only has a faster convergence rate, but also has a smaller steady-state error.