A fast start-up modified LMS algorithm

Abstract

Due to the time-delay spread of the channel, the equalizers for high data rate mobile communication systems has critical importance in mitigation of inter-symbol interference (ISI). One of the most important channel equalization algorithms is the least mean squares (LMS) algorithm. However, the slow convergence rate and the need for long training sequences are the major disadvantages of the LMS algorithm. In this paper, a faststart-up modified LMS (FSU-M-LMS) algorithm based on channel matched filter (CMF) has been proposed to increase the convergence speed and performance of the LMS algorithm. In order to prove the performance of the proposed method, computer simulations are performed on stationary and nonstationary frequency selective Rayleigh fading channels. The obtained simulation results illustrate that the proposed FSU-M-LMS algorithm has better mean square error (MSE), bit error rate (BER) and channel tracking performances than the conventional LMS, modified LMS (M-LMS), normalized LMS (N-LMS) and conventional RLS algorithm.