A new numerically stable structure for fast RLS adaptive filtering

Author

Yu, Xiao-Hu

Author_Institution

Dept. of Radio Eng., Southeast Univ., Jiang-Su, China

fYear

1990

fDate

3-6 Apr 1990

Firstpage

1409

Abstract

A numerically stable structure is presented in which the conventional fast transversal recursive-least-squares (RLS) algorithms and a mixed time-and-order (MTO) updating procedure are combined to provide instantaneous RLS solution for each time step with O( M) computations, where M is the order of filter. Since the new structure is continuously propagated only over 2M+1 time steps, accumulation of the numerical errors is isolated. Both fixed-point analysis and computer simulations show that this structure is long-term numerically stable. Some improved alternatives are considered

Keywords

adaptive filters; filtering and prediction theory; least squares approximations; MTO updating; computer simulations; fast RLS adaptive filtering; fast transversal recursive-least-squares algorithm; fixed-point analysis; mixed time-and-order updating procedure; numerically stable structure; Adaptive algorithm; Adaptive filters; Application software; Computer errors; Computer simulation; Estimation error; Kalman filters; Resonance light scattering; Statistics; Transversal filters;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1990. ICASSP-90., 1990 International Conference on

Conference_Location

Albuquerque, NM

ISSN

1520-6149

Type

conf

DOI

10.1109/ICASSP.1990.115653

Filename

115653