A Complementary Pair LMS Algorithm for Adaptive Filtering

摘要

This paper presents a new algorithm that can solve the problem ofselecting appropriate update step size in the LMS algorithm. Theproposed algorithm, called a Complemen- tary Pair LMS (CP-LMS)algorithm, consists of two adaptive filters with different updatestep sizes operating n parallel, one filter re-initializing the otherwith the better coefficient estimates whenever possible. This newalgorithm provides the faster con- vergence speed and the smallersteady-state error than those of a single filter with a fixed orvariable step size.