Analysis of Adaptive Step-size Normalised Least Mean Fourth Algorithm for Spline Adaptive Filtering

Abstract

This paper presents the spline adaptive filtering which the least mean fourth algorithm is used for cost function. Structure of spline adaptive filtering is described shortly. Normalised least mean fourth algorithm for minimised cost function can converge to optimum values with the fast convergence rate. Adaptive step-size algorithm is derived with an adaptive averaging algorithm. Experimental results depict that the proposed spline adaptive filtering based on normalized least mean fourth algorithm with adaptive step-size parameter can reduce the estimated error rate compared with the traditional least mean square algorithm that can dramatically converge to the steady-state.