Title :
The filtered error LMS Algorithm
Author :
Shaffer, Shmuel ; Williams, Charles S.
Author_Institution :
Diasonics Inc., Milpitas, Ca.
Abstract :
In this paper, we show that the LMS algorithm fails to converge to an optimal solution in the presence of nonstationary noise. To overcome this problem, we introduce the filtered error LMS (FELMS) algorithm. The FELMS algorithm consists of two interconnected adaptive processes. The first process filters the error used by the other. Simulation results demonstrate the superiority of the FELMS configuration with respect to the regular LMS configuration.
Keywords :
Adaptive algorithm; Adaptive equalizers; Autocorrelation; Delay; Filters; Information systems; Intersymbol interference; Least squares approximation; Signal to noise ratio; Transfer functions;
Conference_Titel :
Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP '83.
DOI :
10.1109/ICASSP.1983.1172030
