Event-Driven Adaptive Optimal Control of a Class of Unknown Nonlinear Systems Via Generalized Fuzzy Hyperbolic Models

This work proposes a distributed control architecture for unknown nonlinear continuous-time non-zero-sum (NZS) differential games, integrating event-triggered mechanisms with adaptive dynamic programming. First, a dynamics identifier based on generalized fuzzy hyperbolic models (GFHMs) is constructed using solely input-output measurements, eliminating the need for prior system knowledge. Within the eventtriggered framework, we derive coupled strategy equations for NZS games and implement an identifier-critic structure, where a single critic network estimates near-optimal control policies. A dynamic triggering mechanism, governed by real-time state variations, ensures closed-loop stability in hybrid continuousevent systems and prevents Zeno phenomena without requiring global communication. Numerical experiments validate the method's effectiveness in reducing communication burdens and maintaining control performance.