An improved variable tap-length algorithm with adaptive parameters

Convergence rate and anti-interference ability are significant rules to estimate the variable tap-length algorithm. Moreover conciseness and robustness are important essentials for these algorithms to be utilized in actual application. Following this way, the improved fractional variable tap-length adaptive algorithm that contains variable error width based on fragment-full error and selects the proportion of restricted function by threshold is presented. The proposed algorithm is able to increase the efficiency of the adaptation of error width. Besides, it has specific and reasonable criteria to select restricted factor automatically in different stage. It can improve convergence speed and constrain fluctuation at the same time. The parameters are able to modify reasonably on the basis of different situations. Simulations demonstrate that the proposed algorithm is the nearest to the optimal tap-length on the basis of maintaining good performance. (C) 2017 Elsevier Inc. All rights reserved.