INCHEON — A field-oriented AI maintenance technology developed by Incheon Transit Corp. that cuts the time needed to locate technical manuals during train breakdowns by 90% and saved hundreds of millions of won in outside contracting costs has taken first place in an Incheon city AI contest.
The corporation said on the 13th that its AI Mobile Maintenance Platform, known as Vehicle Doctor, developed in-house by its own employees, won the grand prize at the 2026 Incheon AI Idea Contest.
The contest was open to the city's main offices, its counties and districts, and affiliated corporations and agencies. Vehicle Doctor, which came out on top, began as an idea from Ha Tae-kwan, a manager on the Unyeon light maintenance team at the Unyeon rolling stock depot.
The core of the platform is rapid response in the field. It uses a chatbot-style interface that lets maintenance workers converse with the device and immediately find the recovery manual they need. Compared with the previous method, the platform cuts the time spent searching maintenance documents by up to 90%, improving initial response and the speed of restoring normal operations when a train malfunction occurs, according to the corporation.
Given the security requirements of railway systems, which are designated critical national infrastructure, the platform combines a closed-loop on-device AI that runs without any connection to external networks with AES-256 encryption. The approach eliminates concerns about data leaks, and completing the development in-house without outside contractors saved about 287 million won in budget.
"We applied AI technology to solve the maintenance delays we ran into in day-to-day field work," Ha said. "We will improve maintenance accuracy to support safe urban rail operations."