Revolution in traffic: Paderborn researchers optimize autonomous vehicles!

Revolution in traffic: Paderborn researchers optimize autonomous vehicles!

The University of Paderborn has launched a new research project that aims to significantly improve the interaction between autonomous vehicles and pedestrians. Led by Dr.-Ing. Sandra Gausemeier and Dr. rer. medic. Tim Lehmann, the project focuses on recognizing pedestrians' intentions to act before they actually act. This is intended to help proactively avoid critical traffic situations, which could represent a significant advance in the field of autonomous driving. According to the University of Paderborn, the project was awarded the university's research prize, which is worth 150,000 euros.

The innovative project combines artificial intelligence (AI) methods with movement analysis. Scientists carry out experimental studies on people's decision-making behavior in order to develop predictive algorithms. The challenge lies in real-time generation, processing and response to information, especially in complex urban scenarios.

Goals and methods of the project

One of the main goals is to develop an AI-based system that creates risk profiles and can accurately assess pedestrians' future action intentions. In order to ensure successful pattern recognition of human movement sequences, the quality of the training data is of crucial importance. Therefore, various data collection methods, such as eye tracking and mobile electroencephalography, are used.

A particular focus is on the implementation of experimental studies in urban environments. These tests are intended to help improve and understand human-machine interactions. In addition, it is planned that after training, the autonomous systems will be able to recognize the intentions of pedestrians based solely on onboard camera images. The first results of the project are expected in early 2027.

Technological challenges and security strategies

The research work faces a variety of technological challenges. One of them is the need to optimize autonomous vehicles for normal road traffic in difficult conditions. The Fraunhofer Institute for Cognitive Systems IKS emphasizes that autonomous vehicles work well in test situations, but must be fail-safe in real environments, such as in bad weather or failed sensors. A resilient, intelligent software architecture is therefore sought to ensure the reliability of the systems.

A significant part of the work is also carried out on the KARLI project, which is supported by a consortium made up of the Fraunhofer Institute IOSB, IAO and several industrial partners. KARLI, which stands for artificial intelligence for adaptive, responsive and level-compliant interaction in the vehicle of the future, focuses on AI functions for automation levels 2 to 4. Individual interactions between humans and AI play a central role here. Adapting interactions to different levels of automation is intended to improve driver safety and attentiveness.

Data protection and transparency are also crucial factors in gaining user trust in the technology. Innovative approaches, such as the use of AI-supported sensors, interior cameras and large language models, are intended to ensure that interactions in the vehicle are optimally designed while at the same time the anonymity of the occupants is maintained. The first functions from these projects could be available in series vehicles by 2026.

Overall, the research results from Paderborn and the Fraunhofer Institutes show that the path to safe and effective autonomous driving involves many complex challenges that must be overcome through innovative approaches and interdisciplinary collaboration.