11.–12.06.2025 #polismobility

EN Icon Pfeil Icon Pfeil
EN Element 13300 Element 12300 DE

Artificial intelligence for new mobility concepts

Share page
PrintPrint page Read duration ca. 0 minutes

How does the use of artificial intelligence-based technology impact the transportation transition?

Automated driving

Automated driving

In the field of machine learning (ML) and artificial intelligence (AI), a lot has happened in Germany in the past year, as the follow-up study to the "IDG Study - Machine Learning / Deep Learning 2019" indicates. Once again, companies of different sizes were asked about their assessment regarding ML and AI. Compared to 2019, the number of companies involved in machine learning (ML) increased by 20 percent to almost 73 percent. German companies have thus recognized the importance of artificial intelligence and machine learning, according to the study's summary.

The range of applications for future technologies is wide: AI and ML are finding their ways into industry, energy supply, transportation, logistics and traffic, as well as healthcare and almost all other sectors.

The top four positions of AI and ML applications in use are occupied by speech analysis (51 percent), followed by image recognition, text analysis and translation of texts (with 46 percent each), the study reveals. However, the type of application varies according to company size. For example, large companies tend to rely more on machine learning to capture large document inventories and enhance planning systems, and use ML for facial recognition (52 percent), optical character recognition (56 percent) and planning systems (54 percent).

Traffic control and autonomous driving

Facial and optical character recognition play a decisive role in the traffic sector and in traffic control on the one hand and in the development of autonomous vehicles on the other. Intelligent cameras, for example, observe traffic on the roads as part of assistance systems and perform "semantic segmentation": They recognize objects, traffic signs and other road users and distinguish them from one another, as Professor Anton Kummert from the Chair of General Electrical Engineering and Theoretical Communications Engineering at the University of Wuppertal reports. AI and ML are key technologies when it comes to mobility concepts of the future.

Professor Anton Kummert is one of the founding directors of the interdisciplinary Center for "Machine Learning and Data Analytics" (IZMD), located on campus. With this institution, Wuppertal University is pursuing the goal of creating a cross-faculty institution for interdisciplinary research and transfer in the field of artificial intelligence, machine learning and data analytics. The IZMD has two supporting pillars: scientific research in the aforementioned areas as well as transfer activities and cooperation with regional industry, civil society, public institutions and intermediaries. The transfer pillar is called "Bergische Innovationsplattform für Künstliche Intelligenz (BIT)" and is advised by a transfer advisory board. By providing advice and creating application-oriented content and formats as well as subject-related internships, there is also an exchange with young scientists in the relevant courses of study. However, it is not only about research, but also explicitly about technology transfer with partners in industry, according to Prof. Kummert.

Safety is the dealbreaker

When it comes to automated driving and traffic control, the hopes of the business, politics and urban planners rest on the results of research institutions and research projects like the Wuppertal example. The traffic of the future should not only become smarter - but also safer. The implementation of innovative mobility concepts involving automated components stands and falls with safety. The aspect of safety is essential, for example, with regard to semantic segmentation. While humans often make intuitive assessments and react on the basis of these - for example, when a ball rolls onto the road - the artificial intelligence behind them can only act according to programmed algorithms. This requires the simulation of a comprehensive catalog of scenarios in order to avoid accidents.

Accident prediction models for automated driving

"Automated and connected vehicles will not become prevalent in all sub-areas of the city for a long time. As a result, previously assumed effects - from traffic safety to traffic performance as well as spatial effects - will have to be reevaluated." With this problem in mind, the German Aerospace Center and the Institute of Transportation Systems Engineering in Berlin, funded by the German Federal Ministry of Transport and Digital Infrastructure, are investigating the potential of "artificial intelligence for traffic safety work" in the KI4Safety project.

Central to this is the development of an accident prediction model that predicts accident frequencies as well as their influencing variables at intersections with different geometries. A large amount of image data is automatically analyzed for safety-relevant infrastructure features and patterns and used for the prediction of traffic accidents. Furthermore, the identification of further relevant factors for high predicted accident frequencies is essential. The system will provide support services for practical traffic safety and planning work.

For this purpose, data of environmental factors prevailing at the time of accident events, such as infrastructure data, weather data, traffic engineering parameters and traffic control data are fused with corresponding accident data. With the help of artificial intelligence, accident numbers are estimated from the data and factors influencing accident frequencies. The researchers are testing and optimizing the system in collaboration with police authorities, accident commissions and planners.

Contribution to the traffic turnaround

Expectations regarding a contribution to the mobility transition through the use of artificial intelligence-based technology only apply under certain conditions, as an interdisciplinary study by the Vienna University of Technology shows. The study predicts a reduction in traffic only under the premise that automated vehicles are used as an extension of existing public transport - in other words, if vehicles and journeys are shared. Otherwise, traffic volumes would even increase significantly, according to the authors in Avenue21. Automated and Connected Transport: Developments in Urban Europe.

This open-access publication addresses the impact of automated and connected vehicles on the European city and the conditions under which this technology can make a positive contribution to urban development. The interdisciplinary team of authors from the Faculty of Architecture and Spatial Planning at the Vienna University of Technology highlights two theses that have so far received little attention in scientific discourse: "Automated and connected vehicles will not become established in all subspaces of the city for a long time. As a consequence, previously assumed effects - from traffic safety to traffic performance as well as spatial effects - have to be re-evaluated. To ensure a positive contribution of this technology to the mobility of the future, transport and settlement policy regulations must be further developed. Established territorial, institutional and organizational boundaries need to be challenged in a timely manner."