This page contains automatically translated content.

10/10/2024 | Pressemitteilung

Reducing accident risks for cyclists with AI

Around half of road traffic fatalities worldwide are pedestrians and cyclists, i.e. so-called "Vulnerable Road Users" (VRU). The Department of Communication Technology (ComTec) at the University of Kassel is currently working on the development of VRU protection systems to warn drivers and vulnerable road users of potential collisions and thus prevent accidents.

The photo shows the crossing of a kerb on the test track. The AI measures position and movement data.
Crossing a kerb on the test track. The AI measures position and movement data. Photo: ComTec

ComTec is conducting research into smartphone-based protection of VRUs under the direction of Prof. Dr.-Ing. Klaus David (FB16, Department of Communication Technology). Position and movement data, so-called movement trajectories, of VRUs and vehicles are recorded and exchanged for this purpose. This is done, for example, using the sensor technology of smartphones or wearables and their extensive built-in communication options. The exchanged movement trajectories and AI-based filters can be used to predetermine (predict) overlaps and thus collisions, which in turn can be prevented by means of notifications.

Crossing kerbs is an important event for determining the position of VRUs and predicting collisions. The position of the kerbs is usually known. If the crossing of a kerb can be detected from the movement data of the VRU and combined with the usual satellite-based positioning systems (GNSS), the more precise position of a pedestrian or cyclist can be derived. In addition, crossing the kerb is a reliable indication that the road user is moving onto the carriageway. This is important in order to adapt the collision prediction filters to the situation.

ComTec is researching the detection of kerb crossings by bicycle. For this purpose, movement data is recorded using a bicycle equipped with smartphones/wearables. This data is then used to train and optimize AI algorithms. A test track is needed to obtain the movement data required for training and also for testing over various kerbs commonly installed in Germany. The North Hessian company Profilbeton is sponsoring a test track that can be used to collect such data from cyclists. Various kerbstones manufactured by Profilbeton were installed on the grounds of the University of Kassel for this purpose. The first research results are expected in summer 2025.

The Department of Communication Technology, headed by Prof. Dr. Klaus David, has 20 employees and students researching the collection and processing of context data using AI algorithms, including for VRU protection. The company Profilbeton has been producing special kerbstones and floor indicators (guiding plates for the blind) as well as elements for barrier-free crossings since 1998.

 

Contact:

University of Kassel,
Faculty 16, Department of Communication Technology Wilhelmshöher Allee 73
34121 Kassel
Tel.: 0561 804 6314
Email: comtec@uni-kassel.de

Profilbeton GmbH
Waberner Straße 40
34582 Borken/Hessen
Tel. 05682 7386 0
Email: info@profilbeton.de