Data-driven-based event-triggered optimal control of unknown nonlinear systems with input constraints

This paper is concerned with event-triggered control problem for unknown nonlinear systems with input constraints. By introducing a nominal system and a discounted cost function, the original event-triggered control problem is equivalently transformed into an event-triggered optimal control problem. Then, a data-driven model is designed by recurrent neural networks to approximate the unknown dynamics of the considered system to make the obtained results have wide applicability. After obtaining the system dynamics, a single critic neural network is constructed to acquire an approximate solution of the Hamilton-Jacobi-Bellman equation with multiple nonlinear terms. To achieve the purpose of relaxing the persistence of excitation condition, the update law of the critic NN is designed by using the current data and historical data. By resorting to the Lyapunov stability theory, the proposed event-triggered optimal controller can ensure that the state variables and the critic NN weight errors are bounded. Finally, the effectiveness of the developed control scheme is demonstrated by two simulation examples.