Online Event-Triggered Mechanic for Unknown Nonlinear Continuous-Time Optimal Tracking Systems

This paper delves into the event-triggered (ET) optimal tracking control (OTC) for unknown nonlinear continuous-time (CT) systems. To tackle this challenge, we employ an augmented system that amalgamates the error system dynamics with the reference dynamics, and introduce a discounted factor into the performance index function. Innovatively, we craft an online event-triggered mechanic (OETM) to solve the ET Hamilton-Jacobi-Bellman equation (HJBE). This approach is implemented via an identification-critic framework, featuring two neural networks (NNs): an identifier NN for estimating the unknown system dynamics, and a critic NN dedicated to approximating the solution of the ET HJB. Finally, simulations demonstrate the efficacy of our devised methodology.