Topology-Aware Data-Driven Event-Triggered Control: An Optimization-Based Approach

This paper presents a topology-aware framework for data-driven event-triggered control of unknown continuous-time linear systems. Addressing the conservatism of conventional isotropic triggering conditions, we propose a dual utilization paradigm where offline data are used to synthesize both the feedback controller and a structure-embedded triggering mechanism. We introduce a generalized triggering structure that reshapes the safety region to align with the system’s learned energy landscape. An optimization-based synthesis is formulated to maximize the safety region while strictly guaranteeing global exponential stability and a strictly positive Minimum Inter-Event Time (MIET). Numerical experiments validate that the proposed strategy significantly reduces communication frequency compared to the isotropic baselines.