A controlled block implementation of the NLMS algorithm

The block implementation [1] of the normalized least mean square (NLMS) algorithm has another parameter governing its convergence property. Controlling its block length enables to keep the estimation error at a desired level against the fluctuation of the power ratio of the reference signal to the noise [2]. This block length control technique can be also applied to improving the convergence property. This technique provides almost the same convergence rate as the recursive least square (RLS) algorithm does.