Jakarta, INTI - Fatigue and drowsiness among train drivers are major risks that can trigger train accidents in Indonesia. Conditions within the locomotive cab, such as noise, heat, and dust, combined with non-ergonomic seating, can increase the risk of microsleep, a state where a person unintentionally closes their eyes for more than a second while on duty.
To address this safety challenge, a research team from the Faculty of Industrial Technology (FTI) at Institut Teknologi Bandung (ITB), led by Prof. Ir. Hardianto Iridiastadi, MSIE, Ph.D., CPE, developed a Fatigue Detection Technology system. This system utilizes video cameras to monitor ocular indicators and facial expressions.
Promoting Local Technology Products
Hardianto explained that imported fatigue detection technologies currently on the market are relatively expensive, ranging from hundreds of millions to billions of rupiah. With local development and products, the system can adapt to the specific operational needs of train drivers in Indonesia.
"Many technologies used in Indonesia are imported at very high prices. Yet, our lecturers and researchers are highly capable and able to create devices that are far more affordable, validated, and effective. This research represents our effort to promote the commercialization and adoption of local technology products," said Hardianto on Thursday, August 13, 2026.
How the System Works
The technology developed by the FTI ITB team utilizes a camera installed inside the locomotive cab to record the driver's face and extract various fatigue indicators in real-time.
The system employs algorithms developed in-house by the ITB team to identify approximately 60 facial points. The data is used to precisely measure the eye aspect ratio, the duration of open and closed eye states, and blink frequency down to the millisecond. Micro-changes in facial expressions that are invisible to the naked eye can also be analyzed as early indicators of fatigue.
In addition to camera-based monitoring, the team developed a Psychomotor Vigilance Task device to measure cognitive response time. During testing, operators will see a checkerboard pattern on a screen as a visual stimulus and are asked to respond as quickly as possible. Slower response times indicate higher levels of fatigue or drowsiness.
Technology testing was conducted using a low-fidelity train simulator in the laboratory. Drawing on over a decade of collaboration with PT Kereta Api Indonesia (Persero), the ITB team tested train driver performance and fatigue levels across various journey scenarios, ranging from short routes like the Commuter Line to long-distance trips.
The team still faces several challenges during the downstreaming stage, including software and interface integration as well as fragmented funding. Nevertheless, the process of system iteration and refinement continues.
Looking ahead, the technology is slated to evolve into an integrated prototype comprising a real-time camera, a mini-computer, and an early warning system. The system is expected to transmit data on the driver's condition to an operational control center, allowing fatigue profiles to be mapped on a daily-to-yearly basis and used to help prevent fatal accidents.
Hardianto also emphasized the importance of collaboration between academia, industry, and government in the development of safety technology.
"Research is the foundation of progress. A single workplace accident has a massive impact, resulting in both loss of life and material damage. I hope we can continue to collaborate across universities, industry, and government to produce domestically made technological products that are practical and widely beneficial; this applies not only to the railway sector but also to bus and truck drivers, as well as the oil, gas, and mining sectors," he said.
Conclusion
A research team from the Faculty of Industrial Technology (FTI) at ITB has developed a camera-based Fatigue Detection Technology to detect fatigue and the risk of microsleep among train drivers. The technology analyzes approximately 60 facial points and incorporates a Psychomotor Vigilance Task device to measure cognitive response time. This development serves as an alternative to imported technologies.
Read more: Indonesia’s Cybersecurity Leaders to Convene at IndoSec 2026