Existence and convergence of solutions to the H sampled-data estimation problem

DOI10.1080/002071700445406MaRDI QIDQ4543895zbMATH OpenOpenAlexFDO

Authors Massimo Maroni, Paolo Bolzern, Giuseppe De Nicolao, U. Shaked

Publication date 11 August 2002

Published in International Journal of Control (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1080/002071700445406

zbMATH Keywords

\(H_\infty\) filteringconvergencediscretizationfinite horizon filtersampled-data filteringtime-invariant linear systems

Mathematics Subject Classification ID

(H^infty)-control (93B36) Sampled-data control/observation systems (93C57) Estimation and detection in stochastic control theory (93E10) Filtering in stochastic control theory (93E11)

Review text

The paper has been motivated by an open question about the possibility of ensuring a priori the existence of the sampled-data finite horizon \(H_{\infty}\) filter for a prescribed attenuation level \(\gamma\) without integrating any Riccati equation. The contribution of the paper is twofold. First, a new discretization of the \(H_{\infty}\) sampled-data filtering problem, based on the direct discretization of the Riccati equation with jumps, is proposed. In contrast to known techniques, this discretization is completely general and does not assume the existence of a stabilizing periodic equilibrium. Based on a new discretization, a necessary and sufficient condition and sufficient condition for the finite-horizon filter over a given interval \([0,T_{f}]\) are derived, removing the necessity of integrating any Riccati equation. The second contribution of the paper is the derivation of additional sufficient conditions that ensure the existence of a filter for any value of the horizon length \(T_{f},\) as well as convergence to the infinite horizon stationary estimator when \(T_{f}\) tends to infinity. The proposed conditions introduce suitable regions for the weight on the initial error. By means of these results it is possible to know in advance whether a given filtering problem has a solution, whatever the horizon length is, and whether the finite-horizon filter tends to the stationary one as the horizon length goes to infinity.

Recommendations

• New conditions for the convergence of H/sub ∞/ filters and predictors
• On the sampled-data \(H^\infty\) filtering problem
• Nonlinear \({\mathcal H}_\infty\) filtering of sampled-data systems
• H  filtering for linear continuous-time systems with sampled measurements
• Sampled-data non-linear H infinity estimation and control

Cited in

(3)

• On the sampled-data \(H^\infty\) filtering problem
• Transient and asymptotic analysis of discrete-time \(H_\infty\)-filters
• The dual algebraic Riccati equations and the set of all solutions of the discrete-time Riccati equation

This page was built for publication: Existence and convergence of solutions to the H sampled-data estimation problem

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4543895)