Share Email Print
cover

Proceedings Paper

Online generalized predicitive control combined with a fast transversal filter
Author(s): Suk-Min Moon; Robert L. Clark; Daniel G. Cole
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The concept of generalized predictive control (GPC) design is extended by combining it with the recursive least squares (RLS) system identification algorithm. In this paper, GPC is combined with the classical RLS system identification algorithm, and with the fast transversal filter (FTF), a modified version of the classical RLS algorithm. The classical RLS algorithm is a straightforward approach for identifying a model from input and output data, and one of the advantages of the classical RLS algorithm is that a model is obtained without time-consuming processes like matrix inversion. The FTF is also an RLS algorithm, but it exploits the shifting property of serialized data and thereby results in a substantial reduction in computational complexity. The advantages of both combined algorithms are no prior system model is required, since the process of system identification is performed recursively from real-time system input and output data, and the controller is updated adaptively in the presence of a changing operating environment.

Paper Details

Date Published: 1 August 2003
PDF: 11 pages
Proc. SPIE 5049, Smart Structures and Materials 2003: Modeling, Signal Processing, and Control, (1 August 2003); doi: 10.1117/12.484760
Show Author Affiliations
Suk-Min Moon, Duke Univ. (United States)
Robert L. Clark, Duke Univ. (United States)
Daniel G. Cole, Duke Univ. (United States)


Published in SPIE Proceedings Vol. 5049:
Smart Structures and Materials 2003: Modeling, Signal Processing, and Control
Ralph C. Smith, Editor(s)

© SPIE. Terms of Use
Back to Top