Highly dynamic and fine-grained traffic data acquisition for greater road safety in cities

In many cities, traffic data helps to determine the current local traffic situation. However, the sensor technology used is not suitable for real-time applications, such as warning systems, because the temporal acquisition rate is too low and the local resolution too coarse.

The aim of the hochfein project, which was funded by the German Federal Ministry of Digital and Transport as part of the “mFUND” (Mobilitätsfonds) innovation initiative, was the highly dynamic and fine-granular recording of movement data of road users at particular accident black spots.

In the framework of the project, existing infrastructure such as lighting poles were used for the installation of the developed cameras at an intersection in Berlin. Image analysis methods made it possible to track passing road users along a relevant section of the road. With an AI for predicting future behaviour, it was possible to make predictions on possible conflicts with other road users. In a field test, an assistance system was tested that warns cyclists of conflicts with vehicles turning right via a widget that was displayed on their smartphone.

Besides the planning and implementation of the field tests and the development of the warning strategy, HFC was responsible for the creation of a traffic simulation. This enabled the development and virtual pre-testing of the planned warning system without endangering test persons and before the technical developments provided real data.