Search

We propose a new convex linear parameter varying (LPV) controller design conditions that regionally stabilize LPV discrete-time systems under magnitude and rate saturating actuators. Moreover, the closed-loop system has an ensured exponential stability performance. The design conditions provide regional stabilization in the sense of polyquadratic stability, allowing an estimate of the region of initial conditions yielding trajectories with guaranteed convergence to the origin. A nonlinear first-order model models the magnitude and rate saturation of the actuators. Additionally, we propose a convex optimization procedure to enlarge the estimate of the region of attraction. Our approach is used in real-time nonlinear level control, illustrating the potentialities of the practical application of the method.

International audience; This paper investigates an event-triggered control design approach for discrete-time linear parametervarying (LPV) systems under control constraints. The proposed conditions can simultaneously design a parameter-dependent dynamic output feedback controller and an event generator, ensuring the closedloop system's regional asymptotic stability. Based on the Lyapunov stability theory, these conditions are given in terms of linear matrix inequalities (LMIs). Moreover, using some proposed optimization procedures, it is possible to minimize the number of sensor transmissions, maximize the estimation of the region of attraction of the origin, and incorporate optimal control criteria into the formulation. Through numerical examples, some comparisons with other approaches in the literature evidence the proposed technique's efficacy.

International audience; This paper deals with the local stabilization of continuous-time Takagi-Sugeno (T-S) systems under saturated actuators and driven by sampled-data controllers. The generalized sector condition is used to cope with the saturated signals and the sampled-data closed-loop dynamics is rewritten as a T-S system with saturated input with time-varying delays. Then, by selecting a convenient Lyapunov-Krasovskii Functional and the application of some useful overbounding techniques, new LMI-based conditions are proposed for the design of saturated sampled-data Parallel-Distributed-Compensation controllers, together with the estimation of the closed-loop domain of attraction. A simulation example illustrates the effectiveness of this proposal and shows the achieved conservatism improvement regarding to previous related studies.