AI Predicts Train Driver Errors, Detects Track Hazards... Technology Developed to Prevent Railway Accidents

KRRI demonstrates 12 integrated technologies, including human-error prediction and real-time hazard detection
Analyzing workers' locations and biometric signals to prevent accidents... field trials planned

A technology has been developed that uses artificial intelligence (AI) to predict the likelihood of errors by train drivers and track workers and detect hazardous situations at work sites in real time. Unlike conventional approaches that respond after an accident occurs, the technology identifies risk factors in advance to prevent accidents.

The Korea Railroad Research Institute (KRRI) announced on October 27 that it had demonstrated 12 integrated core technologies for preventing railway accidents using AI and digital technologies at its Osong branch in Cheongju, North Chungcheong Province, on October 26.

A non-contact AI-based biometric measurement gate that automatically measures workers' body temperature, heart rate, and alcohol consumption. Courtesy of the Korea Railroad Research Institute (KRRI)

The technologies unveiled include AI video analysis to detect risks involving train drivers and track workers, biometric signal measurement and job suitability assessments, smart safety equipment, AI-based predictions of the probability of human error, and integrated safety management based on digital twins.

The core of the system is to assess accident risks in advance by comprehensively analyzing workers' conditions and their surrounding environment. AI analyzes workers' locations and work environments to assess the likelihood of human error. If an abnormal situation occurs, such as a worker leaving a designated work zone, it is displayed on a digital twin screen.

In particular, the system connects the entire process, from work preparation and risk assessment to work approval, real-time safety monitoring, and response. The technology is significant because it integrates safety technologies developed separately in accordance with actual railway work procedures.

The technologies are being developed as part of a national R&D project led by the Ministry of Land, Infrastructure and Transport. Technologies to analyze, assess, and prevent human error among railway workers are scheduled for development by the end of 2027, while data-based technologies to assess and predict the safety status of railway systems are scheduled for development by the end of 2029.

However, the demonstration was intended to confirm the technologies' technical feasibility; their effectiveness in preventing accidents in actual railway operations has not yet been verified. KRRI plans to expand field trials with railway operators and establish an institutional foundation for applying the technologies to railway safety policies and safety management work.

"By linking railway safety technologies developed separately in accordance with actual work procedures, we have demonstrated the technical feasibility of managing human error and on-site risks in advance," said Park Chan-woo, a principal researcher at KRRI. "We will expand field trials and further develop the technologies so they can be used in actual safety management decision-making."

This content was produced with the assistance of AI translation services.