Prescribed performance event-triggered optimal control of nonlinear multi-input systems

Abstract

This article proposes a prescribed performance reinforcement learning control (PPRLC) based on event-triggered mechanism for nonlinear multi-input systems, where target error is constrained to a bounded set. Firstly, the constrained optimal control problem is reformulated as an unconstrained stationary optimal problem by using prescribed performance function. Then, the event-triggered mechanism (ETM) is integrated to save communication resources and reduce data transmission volume. In order to study the solution of the Hamilton-Jacobi-Bellman equation (HJB), we use a reinforcement learning (RL) algorithm based on the single-critic neural network (NN) and introduce a new adaptive law to update the NN weights. Based on the Lyapunov function, the convergence of weights and the closed-loop stability of the system are confirmed. Finally, the correctness and effectiveness of the method are proved by a numerical simulation example.