Title :
A new family of gradient-based adaptive filtering algorithms with variable step size
Author :
Du, Zhimin ; Wan, Peng ; Pei, Tingrui ; Wu, Weiling
Author_Institution :
Sch. of Inf. Eng., Beijing Univ. of Posts & Telecommun., China
Abstract :
Under a uniform framework, this paper develops a new family of adaptive filtering algorithms, where the wellknown normalized LMS algorithm and normalized constant modulus algorithm (CMA) are included. They all update the weight according to the gradient descent method, but this time a variable and relatively optimal step size is used instead of a constant one. Some application examples are also given to show their efficiencies
Keywords :
adaptive filters; filtering theory; gradient methods; least mean squares methods; adaptive filtering algorithms; constant modulus algorithm; cost ftinctions; gradient descent method; gradient-based adaptive filtering algorithms; normalized CMA; normalized LMS algorithm; optimal step size; recursive least squares; steady-state performance; variable step size; Adaptive filters; Computational complexity; Convergence; Cost function; Equations; Filtering algorithms; Least squares approximation; Least squares methods; Resonance light scattering; Stochastic processes;
Conference_Titel :
Global Telecommunications Conference, 2001. GLOBECOM '01. IEEE
Conference_Location :
San Antonio, TX
Print_ISBN :
0-7803-7206-9
DOI :
10.1109/GLOCOM.2001.965097
