Trams play a key role in sustainable urban mobility. At the same time, their specific driving dynamics – long braking distances, inability to swerve around an obstacle – makes them particularly challenging in complex traffic situations. Despite the drivers‘ high level of experience, accidents involving property damage or even serious injuries occur time and time again. Against this background, the need to develop driver assistance systems specifically for tram operation arose around ten years ago.

With its 3D real-time stereo vision technology, AIT has created the basis for a driver assistance system tailored to trams. In collaboration with Bombardier (now Alstom) and Mission Embedded, ODAS, the world‘s first driver assistance system for collision avoidance in trams, was successfully developed and put into series production for the first time in an entire vehicle fleet in Frankfurt. The system recognises obstacles, assesses the collision risk in real time and warns drivers of potential dangers. In a later development, COMPAS was created, which, in addition to obstacle detection, also integrates speed control (over-speed prevention) as an additional assistance function.

The technology is now in use in numerous cities around the world, including Frankfurt, Zurich, Blackpool, Duisburg, Dresden, Melbourne, Brussels and Essen. Many new tram vehicles from Alstom are equipped with the system, which means that the assistance system is increasingly becoming established as standard technology. The high demand for such systems among tramway operators is proof of their positive impact on reducing collisions and property damage.

In addition to the safety benefits, the technology also has an economic impact. The developments financed by the vehicle manufacturer enabled long-term value creation: by securing the location and creating jobs in Vienna, particularly at Alstom and Mission Embedded.

The developments are being continued on an ongoing basis. For example, COMPAS trams can already be used to precisely map the route network, which in turn forms the basis for a highly available system for vehicle self-localisation.

AIT is developing AI systems that will enable vehicles to understand scenes for the first time and thus create a basis for the intelligent tram of the future. The research framework has shown that the sensors of the assistance system, in combination with the localisation system and AI scene understanding, enable effective automated monitoring of the rail vehicle infrastructure.