CZ/SK verze

Could Trains See Danger Coming? Czech Trials and Spain’s Fatal Crash

Could Trains See Danger Coming? Czech Trials and Spain’s Fatal Crash
photo: Tadeáš Bednarz, CC BY-SA 4.0 / Wikimedia Commons/AŽD EDITA 811 111-4 at Dětenice station
02 / 11 / 2026

Czech railway technology supplier AŽD is testing cameras and sensors that help trains recognise obstacles ahead. Could such equipment help prevent a collision like the one that recently killed dozens of people in Spain?

A car stuck at a level crossing, animals on the railway or a fire near the track can force a driver to react quickly. AŽD is testing equipment designed to detect these hazards as part of European research into automated trains.

The company presented its trials at a Europe’s Rail workshop under the FP6 FutuRe project, which focuses on improving regional railway services while keeping costs affordable. Testing takes place on the approximately 23-kilometre Kopidlno–Dolní Bousov line in the Czech Republic. AŽD uses EDITA, a rebuilt diesel railcar fitted with cameras, laser sensors and computers. The railway also carries seasonal passenger services.

How Does a Train Recognise Danger?

The cameras watch the track and its surroundings, while laser sensors measure the position and shape of objects. Together, they give the onboard computer information about what lies ahead. The difficulty is deciding what actually poses a danger. A car parked beside the rails, for example, should not stop a train, but car blocking a crossing should. Smoke drifting from nearby land may be harmless to passing trains, while a fire on the track requires a different response.

AŽD tested these situations, including cars obstructing level crossings with and without barriers. Engineers also worked with a farmer whose horses regularly cross the railway, allowing them to test the equipment around animals.

Keeping Track of the Train When Data Is Lost

The company also tested how a train can continue estimating its location when supporting data for satellite positioning becomes unavailable.

In that situation, the system uses other onboard sensors to follow the train’s movement. Its estimate becomes less precise until the missing data returns. The researchers deliberately interrupted the data during testing and reported that the system continued working with performance they considered acceptable for its intended use. That does not mean the train can continue indefinitely without reliable information, but the equipment controlling it must account for the increasing uncertainty about its position.

The work presented by AŽD also covered automated request stops. Passengers would request a stop from inside the train or through a station touchscreen. Requests arriving too late for a safe stop would not be accepted.

These technologies are being developed for regional routes, where operators need affordable equipment that can handle everyday situations before more driving tasks are automated.

Could It Have Helped in Spain?

On 28 October, a Renfe AVE and an Iryo train collided near Urda on Spain’s Madrid–Andalusia high-speed route, as RAILTARGET previously reported. Reuters reported at least 68 deaths. Citing Transport Minister Óscar Puente, it said a communications failure had stopped traffic before the Renfe train was erroneously authorised to proceed. The circumstances remain under investigation.

In theory, onboard sensors could have given the approaching train an additional warning, but at high speed, even detecting another train ahead may not leave enough time to stop. AŽD’s regional trials have not demonstrated whether its equipment could provide that warning early enough in such conditions.

Similar technology is already being tested elsewhere in Europe. In the Netherlands, Alstom and its partners reported detecting obstacles up to 1,000 metres away using equipment fitted to a freight locomotive. Those tests concerned shunting operations, however, rather than high-speed passenger services.

More recent tests in Germany have also linked obstacle detection to automatic braking. RAILTARGET previously reported on the AutomatedTrain project, completed in September 2026, a Siemens Mireo ran without a driver at a test centre and braked automatically when obstacles appeared.

A second train tested obstacle detection on Stuttgart’s S-Bahn network, with a driver remaining in control. The automated journeys were designed for speeds of up to 40 km/h without passengers, so the results do not yet show how the equipment would perform on a high-speed service.

In France, SNCF Réseau is developing MARS LGV, autonomous vehicles intended to inspect high-speed lines before services begin. Cameras, radar and laser sensors would identify obstacles and other problems along the route. This would help check that the line is clear before trains enter service, but would not provide continuous protection from obstacles appearing later.

Testing whether onboard detection could also help protect high-speed services would be a useful next step. An extra warning could give drivers or automatic systems more time to brake, potentially avoiding a collision or reducing its severity. The challenge is making that warning reliable and early enough to matter.

Tags