With the 2019 update of the EU Directive 2019/1936, a systematic road safety assessment will become mandatory for all EU states. The safety level of existing roads is to be improved by investing funds specifically in road sections with the highest accident frequency and the greatest accident prevention potential. This requires not only systematic recording of accidents but also suitable data of the road infrastructure and potential deficits. The collection of this data is very time-consuming and is currently only carried out on an ad hoc basis and on the object level rather than the network level.
In line with the EU directive, the KISStra project aims to investigate the feasibility of a systematic network-wide road safety assessment. The methodology is developed based on artificial intelligence (AI) and validated it with a prototypical implementation. The primary technical goal is to develop the foundations for a universal AI-based methodology for the detection, acquisition and provision of selected safety-relevant indicators from images of the road.
In the project, the requirements for a digital safety audit are investigated and the feasibility in principle is tested by means of a prototypical implementation. In a first step, the AI will be trained for the detection of certain indicators (e.g. traffic signs, lane markings, line guidance, stopping sight distance, etc.). As image data source, the roadway and roadway surface images acquired every 4 years in the context of the condition survey and assessment (ZEB) are to be used. Finally, a rule-based approach will be used to identify possible safety deficits on the basis of the indicators identified.
The project KISStra is realized as a joint project by the HELLER Ingenieurgesellschaft mbH and the Institute of Numerical Methods and Informatics in Civil Engineering of the Technical University of Darmstadt.
