Control Constrained Distributed Event-Triggered Adaptive Control for Dynamically Unknown Large-Scale Systems

This article presents an innovative decentralized control strategy tailored for a class of continuous nonlinear interconnected large systems with unknown internal dynamics. This strategy adopts an online learning optimal control method grounded in neural networks, leveraging neural networks as the cornerstone of online learning optimal control. Firstly, we design a neural network-based identifier to capture and reconstruct the complex and varied internal dynamics of this interconnected system. To approximate the cost function more precisely, we construct a suite of critic neural networks. Through these networks, we can derive the control policy more effectively. Furthermore, we rigorously prove that the dynamics of the estimation error pertaining to the critic networks are stable and ultimately bounded, thereby enhancing the reliability of the control strategy. Lastly, to validate the effectiveness of our proposed decentralized controller, we conduct simulation experiments and achieve satisfactory results.