We have already reported on the theory and field tests of how AI could optimise road traffic on //next. Now, in Hamm, Westphalia, a traffic light has gone into operation that is completely controlled by an AI. It is unique in Germany so far, but there are already partial applications in Munich.
There are many projects for better regulation of traffic flows, reports heise.de for example: The Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB is working on a system in which machine learning is to optimise the flow of traffic at intersections. In the general development towards autonomous cars, work is being done to network the vehicles with the traffic light systems. In Munich, traffic lights have recently started communicating with cars at 22 intersections. There, drivers are shown what speed they have to choose in order to get through an intersection when the light turns green. In London, pedestrian lights are switched to green by default: They only turn red when a vehicle approaches.
In Hamm, the technology of the Munich-based company Yunex, formerly a division of Siemens Mobility, is now being used: "In its 'awareAI' system, it uses video cameras whose images are evaluated by algorithms. Seven cameras have been installed at the intersection for this purpose. These film diagonally across the streets, also recording cycle paths and footpaths."
For example, the system recognises a cyclist 70 metres before the intersection and takes him into account in its calculations for the duration of the green phases. The AI can also recognise whether individual pedestrians or entire groups such as school classes are approaching, the tech portal quotes Christian Breßler, head of traffic technology and control in the civil engineering and green spaces department of the city of Hamm. Accordingly, the green phase is extended so that the entire group can cross the streets. The road users are only recorded for a short time for the control system, nothing is stored, and data protection is guaranteed.
Read more details about this groundbreaking installation here (in German):
Text: Ingo Schenk