For more than three decades, IXPERTA has been connecting the worlds of enterprise communication, infrastructure, cybersecurity, and software. Our footprint extends beyond technology delivery into advanced research and innovation. In the past three years, we have participated in two pioneering projects focused on railway safety. One of the key outcomes is a functional prototype of a device designed to detect obstacles on railway tracks.
Project Profile
- Research and development of a functional railway car model capable of data collection and software integration.
- Development of a simulator to generate training data for obstacle detection in simulated conditions.
Project Goal:
To develop a functional model of a railway vehicle capable of detecting obstacles on the track using hardware sensors, advanced data processing architecture, and artificial intelligence for final identification and driver notification.
Financial Support
The project was co-funded with state support from the Technology Agency of the Czech Republic under the TREND 1 program – supporting industrial research and experimental development.
Project Partner:
IT4Innovations National Supercomputing Center, VŠB – Technical University of Ostrava
Project Duration:
January 1, 2020 – December 31, 2022
Project Output: Functional Detection Unit Prototype
The detection unit is a device that monitors obstacles in the train’s path, serving as a driver assistance interface. It alerts the driver both visually and acoustically.
- Examples of obstacles: vehicles, people, or animals.
- The system can also recognize railway signage. A specially trained neural network identifies signal colors and alerts the driver in case of violations. It also detects speed signs and provides timely notifications.
Capabilities of the functional unit:
- Size of detected obstacle 50 × 50 × 50 cm
- Distance 200–500 m depending on object size
- Day & night operation
- Speed up to 90 km/h (not yet tested at higher speeds)
- Accuracy 97 %
How the Detection Unit Works
- Sensor fusion: RGB cameras, thermal cameras, and LIDAR data combined into a single output.
- AI processing: A custom-designed neural network distinguishes common railway infrastructure from real obstacles. Minor objects such as flying debris are ignored.
- Installation: The device is mounted on the front of the locomotive at approx. 130 cm height, fixed with an ALU frame and magnetic holders for flexibility.
- Testing: Conducted on inspection vehicles, shunting locomotives, and freight locomotives. Future versions can be integrated directly into train construction.
Detection Layer
3 RGB cameras
for obstacle detection
Thermal cameras
for low-visibility recognition
Wide-range LIDAR
for object sizing and distance detection
Driver Interface
Red display
in the driver’s cab
Processing unit
connected to the display
Power Systems
DC converters
Network components
Technological Challenges
Over three years, IXPERTA’s development team created progressively advanced prototypes while solving key challenges:
Vibrations
significant locomotive vibrations required stabilization via silent blocks.
Curved tracks
detecting obstacles beyond standard camera view.
Hardware resilience
ensuring durability in demanding conditions.
Real-time processing
handling large volumes of data without delay.
The project utilized state-of-the-art global technologies, particularly advancements in LIDAR sensors, point cloud data processing, and AI computation modules (JETSON NX, ORIN).
Collaboration with IT4Innovations
IXPERTA partnered with IT4Innovations National Supercomputing Center, part of VŠB – Technical University of Ostrava. IT4I provided a 3D virtual simulator environment and generated extensive datasets to train the detection algorithm.
Using high-performance computing, IT4I recreated realistic railway conditions, including terrain elevation, vegetation patterns, speed profiles, and static/dynamic objects — accelerating development through supercomputer-based simulations.
Conclusion
Railway safety presents unique challenges. A freight train weighing 300 tons requires a braking distance of 2–3 km, while the detection unit can identify obstacles up to 500 m away. Although it cannot prevent every incident, it can alert the driver in time to reduce the severity of accidents.
Test runs confirmed the unit’s ability to detect people, animals, vehicles, and other obstacles, making it suitable as an auxiliary device across various railway vehicles. In the future, its use may extend to tram or trolleybus transport.
The project is currently at the functional prototype stage. Moving forward to prototype and full product phases requires further investment. IXPERTA is actively seeking partners to advance development and bring this technology into real-world deployment.
Collaboration
About IT4Innovations
IT4Innovations National Supercomputing Center, part of VŠB – Technical University of Ostrava, is a leading R&D hub in high-performance computing, data analysis, artificial intelligence, and industrial applications. Since 2013, IT4Innovations has operated the most powerful supercomputers in the Czech Republic, serving both academic and industry research worldwide.
About IXPERTA
For over 30 years, IXPERTA has been a trusted IT partner in the Czech and Slovak markets. We specialize in infrastructure solutions, cybersecurity, custom software development, and artificial intelligence. Our technology team has delivered unique and complex projects, including the Slovak national supercomputer Devana. In recent years, IXPERTA has also contributed to research projects in railway safety, confirming our commitment to innovation and advanced technology.
Published: 4. September 2025
Read also
Get first-hand technology news
Trends, practical inspiration and recommendations in your inbox