Dynamic Event‐Triggered Optimal Control for Unknown Nonlinear Systems Under Control Input Saturation

In this paper, a novel dynamic event‐triggered optimal control scheme for unknown non‐affine nonlinear systems with control constraints is proposed. First, an identifier neural network is constructed using measurable data to approximate the unknown dynamics. Next, a single critic neural network is designed to derive the event‐triggered Hamilton–Jacobi–Bellman (HJB) equation, with the critic weights updated by relaxing the restrictions on initial admissible control and the persistence of excitation (PE) condition. Subsequently, an optimal constrained event‐triggered control (OCETC) scheme is developed, which samples the state and updates the control input when the triggering condition is violated. Furthermore, a dynamic triggering rule with an additional term is defined by introducing a dynamical variable, significantly reducing computational burden. It is proven that the closed‐loop system state, the estimation errors of the critic network weights, and the dynamical variable are uniformly ultimately bounded (UUB). And the Zeno behavior is also excluded. Finally, simulation examples demonstrate the effectiveness of the proposed method.