​Abstract: Modern airports are complex Cyber-Physical Systems (CPS), where effective coordination between physical entities – like pedestrians and Autonomous Guided Vehicles (AGVs) – and computational components is essential for safety and efficiency. This study introduces a novel CPS architecture that integrates Artificial Intelligence (AI) and Visible Light Communication (VLC) to optimize mobility and enhance real-time responsiveness. Using tetrachromatic LED luminaires modulated via On-Off Keying (OOK) and amorphous SiC optical receivers in a mesh-based hybrid topology, the system creates a VLC infrastructure that delivers real-time, location-aware navigation. A custom protocol ensures low-latency, reliable data exchange between agents and the digital core. VLC receivers capture continuous data on agent positions and movements, which is processed by Deep Reinforcement Learning (DRL) agents trained via Q-learning. These agents adaptively manage traffic flow, minimize congestion, and improve throughput. Simulations and experiments confirm the system’s advantages over traditional methods, enabling GPS-independent indoor navigation, efficient mixed traffic coordination, and scalable deployment within smart airport environments.

Keywords: Cyber-physical systems (CPS), Visible light communication (VLC), Internet of things (IoT), Deep reinforcement learning (DRL), Autonomous guided vehicles (AGVs).

_________________________________________________________________________________________