Project ACCorDCopyright: © ika RWTH
They are located at Campus Melaten, the B 56 highway near Aldenhoven, and the Jackerath intersection: The sensors of the ACCorD project coordinated by the Institute for Automotive Engineering (ika) at RWTH record the movements of cars, bicycles, and pedestrians 24 hours a day. "We need as much data as possible to ensure the safety of automated driving," explains ika employee Laurent Klöker. "In mixed traffic, as it will develop in the coming years, the automated vehicle must be able to react optimally to the behavior of other, human road users."
Since just about anything can happen in road traffic, from sharp steering to sudden, unwarranted lane changes to spontaneous emergency braking, only a vast amount of data can prepare the supposedly intelligent machine on the road. "The sensors record situations such as overtaking, merging, and many others," Klöker says. "We're particularly interested in situations where things have gotten a little risky." That's because researchers can learn the most from them. With this in mind, the three test sites supervised by RWTH were carefully selected: The highway section covers two acceleration lanes, the highway a lot of traffic – including agricultural traffic – and the campus area has many bicycles, pedestrians, and buses. For data protection reasons, the recorded images are automatically processed locally at the measuring points, evaluated there, and immediately deleted again, as Klöker affirms.
"Unfortunately, the project has been repeatedly delayed due to the pandemic, the global chip shortage, and the flood disaster and the related shortage of construction materials," Klöker says. "Funding from the German Federal Ministry for Digital and Transport has now expired, but we hope to receive follow-up funding soon."