Incheon subway to deploy AI inspection robots on tracks by 2027

Incheon Transit Corp. is accelerating the introduction of AI-based autonomous inspection robots to improve the safety and maintenance efficiency of its urban rail civil and track infrastructure.

According to the corporation, the AI-based autonomous inspection robot development project has been underway since March 2025 in a profit-sharing partnership with A2Mind Co. The project aims to reduce human error and the risk of safety accidents — key limitations of the current labor-intensive inspection approach.

The robots developed so far come equipped with core autonomous functions including self-guided and remote operation, obstacle and switch recognition, positioning, and communication-failure response. Rail wear measurement and fastener and facility inspection capabilities have also been fully implemented.

A deep learning-based anomaly detection AI independently diagnoses problems such as rail fastener defects and cracks in ballast and sleepers.

The corporation plans to continuously train the system on field data to maximize the AI's inspection accuracy.

Real-world validation is also well underway. Since August, the corporation has been conducting three months of test runs across the full 165.128 km of track, covering Incheon Lines 1 and 2 and the Incheon-Bucheon section of Line 7.

During this process, the robots' driving stability, communication quality and structural interference are being comprehensively verified, and additional training data is being collected.

Based on the test-run data, the corporation plans to refine the remote-control environment and improve performance before formally deploying the inspection robots in the field starting in May 2027.

Inspection data collected by the robots will later be integrated with a smart urban rail maintenance system being co-developed with the Korea Railroad Research Institute, enabling precise data-driven infrastructure management.

"This project is an innovative challenge to apply cutting-edge AI and robotics technology directly to urban rail operations," Incheon Transit Corp. President Choe Jeong-gyu said. "Through rigorous field verification, we will fully elevate the autonomous inspection capability, improve the accuracy of railway facility inspections and do our utmost to ensure public safety."