A family of adaptive filtering algorithms based on the variable step size

Abstract

In this paper, we present the most used adaptive filtering algorithms such as Least Mean Square (LMS) and its normalized version NLMS with their advantages and drawbacks, and then show how the Variable Step Size (VSS) algorithms have been proposed to slove problems coming from the fixed step size. Series of simulations have been carried out under different effects such as: the size of the adaptive filter and different values of step size to validate the good behaviour of the four presented VSS based algorithms over the classical adaptive filtering algorithms with fixed step size. Also, results have confirmed the superiority of VSS based algorithms in terms of convergence speed with almost identical computational complexity.