Enhanced LMI conditions for observer-based $H_\infty$ stabilization of Lipschitz discrete-time systems
Résumé
This paper deals with observer-based controller design for a class of discrete-time systems with Lipschitz nonlinearities. Usually, the observer-based control synthesis for the considered class of systems leads to the feasibility of a Bilinear Matrix Inequality (BMI). Since, solving a BMI constraint has been an NP-hard optimization problem, then linearizing this constraint to get a convex one is an interesting issue because Linear Matrix Inequalities (LMIs) are easily tractable by numerical softwares (LMI Toolboxes,..). Hence, the aim of this paper is to develop a new Linear Matrix Inequality (LMI) condition, ensuring the asymptotic convergence of the observer-based controller. Due to the introduction of a slack variable technique, the usual BMI problem is equivalently transformed to a more suitable one, which leads to less conservative and more general LMI condition compared to the existing methods in the literature. Conjointly to the slack variable technique, the Lipschitz property and the Young’s relation are used in a reformulated way to obtain additional decision variables in the LMI. In the aim to further relax the proposed LMI methodology, sliding windows of delayed states and measurements are included in the structures of the controller and the observer, respectively. The obtained LMI is more general and less conservative than the first one, which can be viewed as a particular solution. To show the effectiveness and superiority of the proposed methodology, some numerical examples and comparisons are provided.