RLS estimation of input/output models for distributed systems in the presence of noise

Author

Gibson, J.S. ; Lee, G.H. ; Wu, C.-F.

Author_Institution

California Univ., Los Angeles, CA, USA

Volume

2

fYear

1997

fDate

10-12 Dec 1997

Firstpage

1672

Abstract

This paper discusses recursive-least-squares (RLS) estimation of parameters in digital input/output models of linear time-invariant distributed systems. The equivalence between the parameter estimation problem for infinitely long data sequences and a linear-quadratic optimal control problem on a finite interval is used to compute theoretical asymptotic values for the parameters estimated from finite data sequences. Numerical results are given for a sampled-data version of a wave equation

Keywords

distributed parameter systems; least squares approximations; linear quadratic control; multidimensional systems; noise; recursive estimation; sampled data systems; wave equations; distributed parameter systems; infinite dimensional systems; input/output models; linear time-invariant systems; linear-quadratic optimal control; noise; parameter estimation; recursive-least-squares; sampled-data systems; wave equation; Estimation theory; Hilbert space; Lattices; Optimal control; Parameter estimation; Partial differential equations; Resonance light scattering; System identification; Transfer functions; White noise;

fLanguage

English

Publisher

ieee

Conference_Titel

Decision and Control, 1997., Proceedings of the 36th IEEE Conference on

Conference_Location

San Diego, CA

ISSN

0191-2216

Print_ISBN

0-7803-4187-2

Type

conf

DOI

10.1109/CDC.1997.657759

Filename

657759