Robust control of set-valued discrete-time dynamical systems

DOI10.1109/9.654887MaRDI QIDQ4395813zbMATH OpenOpenAlexFDO

Authors John S. Baras, Nital S. Patel

Publication date 12 August 1998

Published in IEEE Transactions on Automatic Control (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1109/9.654887, http://hdl.handle.net/1903/5546

zbMATH Keywords

differential gamesdiscrete control systemdynamic inclusiondynamic programmingfeedbackrobust controlstabilityuncertainties

Mathematics Subject Classification ID

Dynamic programming in optimal control and differential games (49L20) Feedback control (93B52) Nonlinear systems in control theory (93C10) Discrete-time control/observation systems (93C55) Perturbations in control/observation systems (93C73)

Review text

The authors consider a finite dimensional nonlinear control system governed by a set-valued discrete inclusion. The problem setting is derived from a control problem for a discrete dynamic system subject to persistent bounded noise. It is assumed that the right-hand sides of the systems characterizing the intrinsic dynamics of the system and the observer are compact-valued multifunctions. It is assumed that the zero steady state solution is one of the trajectories generated by the trivial control. Three concepts of stability of a control are introduced. The weak asymptotic stability means that there exists an asymptotically stable selector of the mapping obtained via substitution of the control and of a trajectory into the multi-valued right-hand side. Ultimate stability is defined as the requirement that all trajectories generated by the control converge to a bounded set at infinity. Finite gain stability requires that if the control admits two trajectories, then the \(l_q\)-norm of the difference between respective output signals is estimated through the \(l_q\)-norm of the difference between the trajectories plus a term estimating the deviation of the initial point from the origin. Since the right-hand side of the system is a set-valued function derived from a single-valued system with uncertainties, a control which is stable in all the three senses mentioned above is called robust. A robust control is searched both for finite and infinite horizon problems in two forms: state feedback and output feedback. Necessary and sufficient conditions of robustness are obtained via the method of dynamic programming. The uncertainties are treated as an unknown input and thereby the original control problem is converted into a dynamic game. Techniques similar to those used in the theory of differential games are applied. As an example, a robust control is obtained numerically for a 3D nonlinear system with a linear scalar observation.

Recommendations

• Robust control of discrete time uncertain dynamical systems
• scientific article; zbMATH DE number 434710
• Robust stabilization of uncertain systems

Cited in

(9)

• The asymptotic and oscillatory behavior of neutral difference inclusions.
• Continuous and discrete-time stability of a robust set-valued nested controller
• Output feedback receding horizon control of constrained systems
• Control of constrained nonlinear uncertain discrete-time systemsviarobust controllable sets: a modal interval analysis approach
• Adaptive NN control for a class of discrete-time non-linear systems
• Some recent aspects of differential game theory
• State estimation bias induced by optimization under uncertainty and error cost asymmetry is likely reflected in perception
• An underlying theory of multifunctions in discrete-time set-valued dynamical systems
• Renewal processes under coordination opportunities via robust dynamic coalitional games

This page was built for publication: Robust control of set-valued discrete-time dynamical systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4395813)