The German Aerospace Center (DLR) aims to improve safety during the market ramp-up of autonomous driving with a digital infrastructure, sensors on traffic lights, streetlights and buildings, so-called edge computers, and a project called "Managed Automated Driving" (MAD). Some aspects are already known.

Upgraded intersection

ensured sufficient safety
in warehouse complexes and logistics facilities through additional magnetic strips, reflectors on could thus also be beneficial for significantly faster orientation in public transport. Insights from existing logistics applications could potentially be transferred. However, autonomous driving in public spaces remains a challenge of a completely different order. There, situations are considerably more complex and often more confusing.

Defusing security concerns

Existing vehicle-based automation systems, as has been known from projects in Switzerland for some time, sometimes have to drive very slowly for safety reasons in order to react to hidden obstacles. This can make them uneconomical in densely built-up areas, or they may be perceived as overly cautious and slow. In the "MAD Urban" project (analogously, in the transport sector it would probably be called "MAD Logistics"), a complementary approach was therefore developed in collaboration with companies and another research institution and successfully demonstrated in practice for the first time.

From the feasibility study. Photos: U-Shift MAD / DLR

Edge computers, for example, are intended to help – compact computers that process data directly where it originates, i.e., at the edge of a network. "The data collected by these sensors is then combined to create a comprehensive picture of all road users. This also includes objects that are obscured from the vehicle's perspective. In this way, the autonomous vehicle 'sees' more and can cope better with challenging situations," explains Julian Schindler, DLR researcher and technical lead of the project. "MAD can make autonomous driving safer, faster, and therefore more economically attractive – and thus accelerate its introduction in cities." In the future, the data from the digital infrastructure could also be used to manage traffic more efficiently, for example, by adjusting traffic light timings to the current traffic flow.

Practical test passed

In mid-September 2025, a team from the German Aerospace Center (DLR) and the Research Center for Information Technology (FZI) demonstrated the feasibility of this approach in practice using specially equipped research vehicles – a world first in public road traffic and an important milestone in technological development. The vehicles crossed an intersection in Braunschweig, which, as part of the project, was equipped with two sensor columns . For research purposes, sensors and computers record and analyze the traffic . They capture (in compliance with data protection regulations, of course!) the outlines of passing vehicles, people, and objects, creating a picture of the overall traffic situation. Essentially, it's a kind of "warehouse management system" scaled up to a large urban area. It could also be considered a "digital twin" (editor's note).

Whether car or truck...

As the research vehicles approach the intersection, the Edge computer takes over control, continuously calculating the safe route and monitoring the execution of maneuvers. Once the vehicles leave the intersection, the vehicle's automation system reactivates. "You can think of the MAD system like a pilot in aviation or shipping. MAD provides support for safe and efficient driving in challenging situations, thus increasing safety for all participants," explains Julian Schindler. The trials were conducted in close cooperation with the city of Braunschweig and also utilized the sensor technology already installed at Tostmannplatz as part of the DLR's Lower Saxony test field.

Interoperability required

Currently, there are no binding standards for methods and approaches like MAD. However, interoperability will generally be important, meaning that different systems in vehicles and of the scope of automation between vehicles and infrastructure in the future, depending on the situation and .

One example would be a bus line completely controlled by MAD (Machine Adhesion). This would eliminate the need to equip every bus with expensive autonomous driving technology. If many vehicles are to operate automatically in the future, MAD would also offer an economic advantage. Furthermore, new business opportunities would arise, such as offering infrastructure data as an optional digital service.

No binding standards yet

In the "MAD Urban" project, the German Aerospace Center (DLR) collaborated with renowned companies such as Vitronic, a manufacturer of image processing systems for traffic technology, on the sensor columns, and Balluff, the sensor specialists, on stereo image acquisition. Intel supported the project with the safety architecture for vehicles and infrastructure automation. The IT service provider and software developer GFT contributed to the software development for the control center that monitors the automated vehicles.

www.dlr.de

See also U-Shift feasibility study