Robust dynamic event‐triggered control of saturated nonlinear systems using reinforcement learning

Abstract

AbstractWe develop a robust dynamic event‐triggered control (ETC) law for asymmetrically input‐saturated nonlinear systems having matched disturbances. Initially, by constructing a novel cost function for the associated nominal system, we transform the constrained robust stabilization problem into an unconstrained optimal control problem. Then, we propose a dynamic event‐triggering rule for improving the performance in reducing the computational burden. Based on such a dynamic triggering rule, we present the event‐triggered Hamilton‐Jacobi‐Bellman equation (ET‐HJBE). To solve the ET‐HJBE, we employ a critic approximator with its parameters being updated in the reinforcement learning framework. After that, we apply Lyapunov's direct method to prove uniform ultimate boundedness of the closed‐loop system and the critic approximator's parameter estimation error. Finally, we provide two nonlinear plants to validate the present dynamic ETC law.