Adaptive pilot filtering for LMS algorithm

An adaptive pilot filtering procedure is proposed to increase the convergence speed of the LMS (least mean square) algorithm. The procedure uses an adaptive filter with only a few filter coefficients to filter both the received signal and the original signal for the purpose of whitening the received signal. The convergence speed of the adaptive pilot filtering procedure combined with the LMS algorithm is comparable to that of the Kalman algorithm for some applications. The procedure combined with the main LMS algorithm requires less than twice the computation power of a conventional LMS algorithm.>