Global ports increasingly face prolonged stabilization delays, weather-induced operational bottlenecks, and heavy vendor lock-in when deploying traditional rule-based automation. Motivated by these industry challenges, the Yeosu Gwangyang Port Authority (YGPA) initiated a paradigm shift toward an independent, resilient, and highly adaptive digital infrastructure.
Targeted for completion by 2029, our “Physical AI Port” project aims to construct a 1.36 million TEU automated terminal. The primary objective is to transition from passive, rule-based software to an active “Physical AI” ecosystem. By leveraging a 4-step cognitive loop (Perception, Cognition, Prediction, Execution) and Port-Specialized Foundation Models (SLM/LLM), the system enables direct hardware micro-control, autonomous fleet coordination, and real-time digital twin synchronization.
Currently executing our 2026 Phase 1 infrastructure deployment, YGPA has successfully integrated 3D LiDAR and mmWave IoT sensors to achieve absolute spatial awareness with zero blind spots. We are validating SLAM-based, infrastructure-free navigation and utilizing Deep Reinforcement Learning for swarm AGV dynamic routing. These innovations are projected to slash safety-related accidents, reduce fuel consumption, and lower overall OPEX.
Directly addressing the IAPH World Ports Sustainability Program (WPSP) evaluation criteria, this initiative establishes a new global standard for Digitalization and Resilient Infrastructure. It champions Climate and Energy goals by drastically cutting SOx, NOx, and CO2 emissions through a smart Vehicle Booking System and optimized equipment routing. Furthermore, it drives Community Building by localizing high-tech R&D, creating advanced technology jobs, and ultimately delivering a sustainable, exportable “K-Physical AI Port” model to the global maritime community.