Robust estimation for discrete-time linear systems

Author

Mangoubi, R.S. ; Appleby, Brent D. ; Verghese, George C.

Author_Institution

Div. of Control & Decision Syst., MIT, Cambridge, MA, USA

Volume

1

fYear

1994

fDate

29 June-1 July 1994

Firstpage

656

Abstract

Finite-horizon estimators are derived for time-varying, discrete-time linear systems that are robust to a general class of modeling uncertainties. A game theoretic formulation is used to obtain equations for a class of estimators which include the H_∞ and the H₂ (Kalman filter) estimator, as well as filters which trade off these two norms. The solution is obtained in two stages, with each stage requiring the solution of a Riccati equation. The estimators, which have linear dynamics, guarantee a bound on the induced norm of the mapping from the input noise to the output estimation error for all admissible uncertainties.

Keywords

Kalman filters; Riccati equations; discrete time systems; estimation theory; game theory; linear systems; time-varying systems; H_∞ estimator; H₂ estimator; Kalman filter; Riccati equation; finite-horizon estimators; game theoretic formulation; linear dynamics; robust estimation; time-varying discrete-time linear systems; Aerodynamics; Estimation error; Game theory; Laboratories; Linear systems; Noise robustness; Nonlinear filters; Riccati equations; Robust control; Uncertainty;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, 1994

Print_ISBN

0-7803-1783-1

Type

conf

DOI

10.1109/ACC.1994.751821

Filename

751821