‘Rail is the perfect use case for assistance and autonomous technology’

CAF and OTIV launched a joint project in July to test advanced driver assistance features for trams. For OTIV, a Belgian provider of autonomous technology for rail, the cooperation marks the next development step towards the commercial rollout of advanced driver assistance, company founders Niels Van Damme and Sam De Smet tell RailTech.com.

The test programme is ongoing in the Spanish city of Zaragoza, where in the coming five months the two companies will test intelligent collision avoidance technologies and advanced driver-assistance systems (ADAS) features. CAF and OTIV are carrying out their project as part of the Technologies for the Autonomous Rail Operation (TAURO) programme under the umbrella of the Europe’s Rail Joint Undertaking (ERJU). The local operator Tranvia de Zaragoza is on board as a partner.

“CAF and OTIV have the common goal of ensuring safe and efficient operation of trams. ADAS can contribute greatly in that regard”, De Smet says. “In a first phase, an ADAS system supports drivers in avoiding collisions between the tram and other participants to traffic. Additionally, this technology is one of the essential building blocks for Automatic Train Operation (ATO) in urban environments, which in turn can facilitate cost savings and efficiency gains in urban rail .”

Perception technology

The cooperation came about following a successful project OTIV completed in the Netherlands for Dutch Railways (NS) and infrastructure manager ProRail. There, OTIV focussed on assistance systems for passenger train shunting, deploying its sensor suite for image recognition and sensor fusion. As CAF is collaborating closely with NS on ATO, including regular tests on ATO over ATB, the Belgian start-up appeared on their radar.

“In mobility, there is a general trend towards ADAS and autonomous driving technologies. OEMs and operators are essentially on the lookout for parties with certain expertise in that field. Our expertise is perception technology, which is an important building block for continued development and eventual rollout of ADAS and ATO-related technologies”, adds Van Damme.

Non-controlled environments pose particular challenges

Indeed, OTIV stands out for its focus on technology for railway operations in complex environments. Its perception systems and underlying algorithms are geared towards complex environments, such as trains on large industrial complexes and trams in urban centres. These non-controlled environments pose particular challenges for perception systems, as the operating environments of these railway vehicles are shared with cars, pedestrians, trucks and other traffic.

“The technology has to be able to recognise all potential danger, which requires state-of-the-art machine learning algorithms to create a reliable image of the environment. We also have to consider which sensors (incl. lidar, camera and radar) are most suited for this type of project, and where to install those,” says De Smet.
The project in the urban setting of Zaragoza is also different from the company’s earlier projects in that this new environment opens up the road to additional validation steps of the technology. “This time around, the project is less experimental; the tram is running in regular operations , collecting real-life data. This allows the system to learn from every mile driven”.

Railway vehicles uniquely suited for ADAS and ATO technologies

As any car manufacturer will tell you, getting ADAS to function smoothly and perform predictably, is no small feat. In automotive, the deployment at scale of self-driving technology still has a long way to go, especially in an urban setting. But that is exactly why railway vehicles are so suited for ADAS and ATO technologies, De Smet notes.

“Rail is the perfect use case for assistance and autonomous technology. These vehicles run on tracks, greatly reducing the degree of freedom compared to cars or trucks. In the case of trams, the vehicle can do only two things: stop or move forward. Moreover, at OTIV we’re convinced that these technologies are essential elements in our efforts to increase the modal share of rail. Building the mobility system of the future requires a backbone: a transport mode capable of transporting high volumes (of both people and goods), in a fast, cost-efficient and sustainable manner. Rail is the perfect candidate for this task, on all levels: from an international network of freight rail to the local urban tram network. Thanks to assisted and autonomous solutions, rail will be able to take this role and guarantee cost efficiency, flexibility and safe operations”