On the asymptotic analysis of the smoothed least mean square algorithm and the relation with other LMS-type algorithms

The asymptotic performance of the smoothed least mean square (SLMS) algorithm is investigated. It is found that the convergence speed of the SLMS algorithm is approximately equal to that of LMS algorithm in practical applications. Based on the results obtained, the relation between the SLMS algorithm and other LMS-type algorithms is considered.<
>