The Korea Railroad Corporation (KORAIL) has begun developing an AI-based training simulator for track workers.
KORAIL announced on July 23 that it is working on an 'AI-based track work training simulator' that allows workers to learn safe working methods without entering the tracks.
This development project has been selected as a key initiative in the 'AI Application Product Rapid Commercialization Support Project,' which is jointly promoted by 11 government agencies aiming to elevate South Korea to a leading position in AI. KORAIL plans to commercialize the simulator by 2027.
The total project cost is 4.07 billion won, with 2.85 billion won funded by the government. The project will run from this month until December 2027, spanning a total of 18 months. KORAIL is collaborating with Hannam University and YMX on this initiative.
The simulator will utilize AI, virtual reality, and digital twin technology to recreate actual track environments in a virtual space. Workers will be able to repeatedly practice the procedures for track work from preparation to completion in a risk-free environment.
KORAIL expects that this training will enhance the adaptability and risk response capabilities of workers entering the tracks, and plans to expand the training to include not only new employees but also personnel from partner companies.
The simulator will feature virtual representations of tracks, facilities, and surrounding terrain. An AI physics engine will visualize risk factors at the site, such as wind pressure from train operations and the risk of electric shock from overhead lines, allowing workers to train for various scenarios tailored to their roles and specific sections.
An interactive coaching system will be introduced, with an AI agent acting as a 'virtual instructor.' This system will analyze workers' actions and detect risky behaviors in real-time, suggesting correct working methods.
Given the importance of preventing safety accidents during track work, the approach of using an AI agent in a virtual environment to repeatedly train for real risks is expected to enhance the effectiveness of training before workers are deployed on-site.
KORAIL plans to secure highly realistic training content through the simulator's development and validation process, gradually applying it across railway sites nationwide.
Yang Tae-hoon, head of KORAIL's Railway Research Institute, stated, 'The core of this development is to safely and repeatedly experience the risk factors of the field in a digital space and learn how to respond. We will enhance the safety capabilities of track workers through practical training utilizing AI and digital twin technology.'
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.