Observer-based Event-triggered Robust Control for Input Constrained Uncertain Nonlinear Systems

In this paper, we propose an adaptive dynamic programming (ADP)-based event-triggered robust control (ETRC) method for continuous-time input constrained nonlinear systems with uncertainties. By constructing a neural network (NN) observer, the estimated uncertainties are employed to design a novel value function, which reflects the real-time uncertainties, regulation and control input simultaneously. Then, a critic-only structure is adopted to approximate the value function and obtain the control law. Based on the event-triggered strategy, the control law is updated when events occur. Thus, the computational burden is reduced, and communication resources and bandwidths are saved. Furthermore, we prove that the closed-loop system and the approximate error of the critic NN weights are both uniformly ultimately bounded by using Lyapunov's direct method. Finally, an example is employed to verify the effectiveness of the developed ETRC method.

I. Introduction

Nowadays, unavoidable dynamic uncertainties in engineering control systems, such as aerospace, electrical systems and power systems, may degrade the control performance or even cause the closed-loop system unstable. To address this problem, robust control has drawn intensive attention in the control research field for several years, and many remarkable results have been obtained [1]. As a popular idea, the robust control (RC) problem is usually converted to an optimal control (OC) problem by establishing a modified performance index function. However, it is difficult or impossible to solve Hamilton-Jacobi-Bellman (HJB) equation for complex nonlinear systems. Fortunately, as an intelligent strategy, adaptive dynamic programming (ADP) can overcome this difficulty by forward computation [2].