Adaptive Dynamic Programming-based Self-triggered Optimal Control for Nonzero-sum Games of Nonlinear Systems with Constrained State

Taking into account multiple control inputs problem in constrained states systems, based on the adaptive dynamic programming (ADP) approach, a sort of self-triggered control (STC) method has been proposed. The constraint boundary of the state constrained system is released by introducing a potential barrier function. Then, on the basis of event-triggered control (ETC) law, the STC strategy was obtained. The STC strategy was updated, if the time reached the triggering time calculated by the self-triggered mechanism. Next, the neural networks (NNs) are used to approach the optimal value function, and then an approximate self-triggered optimal policy is designed. To solve the problem of the persistence of excitation (PE) conditions, the experience replay (ER) method is combined with the weight updating rule, and the validity of the proposed method is proved by a example.