If car accidents are a thing of the past at some point, then thanks to an inconspicuous stretch of the main road in Garching. More precisely: thanks to the B 471 between the Garching-Süd motorway exit and the intersection with Zeppelinstrasse. There, scientists from the chair for robotics, artificial intelligence and real-time systems at the Technical University of Munich (TU) are researching a “digital twin” of road traffic.
The project is called Providentia ++. The Latin word means “Providence” and stands for “proactive video-based use of telecommunication technologies in innovative car traffic scenarios” and means: cameras, radar and similarly functioning lidar systems with laser beams collect data in real time that are used to virtually map the traffic situation. Networked or even autonomous vehicles access this digital twin and can thus drive more efficiently and safely. For example, the system recognizes which vehicles are driving in which lanes and how fast, long before the driver sees it. In this way, all road users get through the traffic optimally, which avoids traffic jams. The system also helps to prevent accidents, as the sensors can foresee dangers.
To make this scenario a reality, the Providentia researchers have been testing the system since 2017 on a section of the A 9. Here, high-resolution cameras and radar systems are mounted on two gantries. Andreas Schmitz from the TU assures: “Nobody is flashed here, no license plates or faces are saved. Our project is maximally data protection compliant.” The data is transmitted wirelessly via 5G. Artificial intelligence helps to recognize vehicle types and classes and to merge the data – i.e. to create the said digital twin. The aim of the scientists is to develop services that can, for example, use the information from the digital twin to independently change lanes on the motorway or to slow down.
Providentia ++ is the second phase of the research project. New sensor masts and computers have been in place for two weeks; The first cameras, radar and lidar devices were installed on the sign gantries above the B 471. The 50 new sensors for Providentia ++ are all connected by a network. The first data should flow in a week. The researchers are currently concentrating on making the existing digital twin as reliable as possible, i.e. it can also be used in fog, snow and heavy rain.
After the A9, the B471 in Garching will also become a test track for the Providentia project.
(Photo: Robert Haas)
Artificial intelligence now recognizes itself when measurements are unreliable; if the data suggests, for example, that objects are zigzagging or racing across the sidewalk. The extended test track is predestined for this, as cyclists, pedestrians and different types of vehicles come together here. The intersection of the B 471 and Daimlerstrasse serves in the project as a blueprint for complex intersections in city traffic. Only if the digital twin functions correctly at every level of complexity can it help to make road traffic safer.