Intelligent Road Maintenance and Smart Infrastructure
AI-powered road surface inspection and autonomous repair systems for Jeju smart infrastructure
Overview
Launched in August 2025 as part of the RISE (Regional Innovation System & Education) initiative, this project develops AI-powered systems for detecting road damage and planning maintenance on Jeju Island. The idea is straightforward: instead of waiting for potholes to cause accidents, use computer vision and drones to find problems early and fix them proactively.
Technical Approach
The system combines three key technologies. Edge AI runs lightweight computer vision models directly on roadside devices and vehicle-mounted cameras, enabling real-time damage detection without relying on cloud connectivity. Drone-based imaging provides aerial surveys for comprehensive road condition assessment across Jeju’s extensive road network. Reinforcement learning optimizes maintenance scheduling, learning from historical data to predict where damage is likely to occur next and prioritize repairs accordingly.
Together, these components form an AI-Predictive Maintenance System (AI-PMS) designed for sustainable infrastructure management.
Why Jeju
Jeju Island’s road network faces unique challenges — volcanic soil conditions, heavy tourist traffic, and frequent weather events. The island also serves as an ideal testbed for smart city technologies due to its manageable scale and strong local government support for innovation.
Collaborators
- Cheju Halla University: Lead research institution
- Seoul National University: AI and computer vision research
- KAIST: Advanced robotics and reinforcement learning
- RovoRoad: Industry partner for technology deployment