The machine perception of the driving environment is an interdisciplinary topic at the engineering institutes in F3. It is achieved using various sensors installed in the vehicle, such as cameras (stereo or mono cameras), radar sensors, ultrasonic sensors, and lidar sensors. All of these sensors have specific advantages and disadvantages in terms of their measurement properties and are therefore combined. This is referred to as information fusion. The aim of environment detection is to create a dynamic vehicle environment model that represents other road users and their status and is made available for further processing. It should also include relevant infrastructure such as traffic signs and traffic lights, as well as structuring elements such as traffic islands, curbs, lane markings, and pedestrian crossings. The resulting research topics include sensor data processing for the detection and classification of objects, in particular using machine learning methods, as well as novel tracking and fusion methods for more accurate environment detection in complex urban scenarios.

Building on environment models from environmental detection in the vehicle and in intelligent infrastructure, new algorithms for implementing automatic driving functions are being developed in F3. These include methods for situation recognition based on the relationships between all individual components of the environment model, for calculating a machine understanding of scenes based on their dependencies, and for situation prediction. Based on this, algorithms for higher-level action planning to specify vehicle behavior and strategies for trajectory planning are being investigated, taking comfort and safety into account. In addition to classic methods, machine learning techniques are also increasingly being used in this area of research. The interaction of these modules enables the vehicle to respond to the actions and reactions of other road users. Within the framework of F3, automated test vehicles with approval for public transport are available, which we use to test the methods we have researched.

This area of research deals, among other things, with the question of how connecting vehicles with each other and with the transport infrastructure can make traffic more efficient and safer and our mobility more environmentally friendly. One example is the provision of information about other road users at blind spots or cooperative behavior at intersections to improve traffic flow. In addition to the development of the actual functions, this also raises a number of questions about the scalability, reliability, trustworthiness, and service quality of the communication protocols, as well as new requirements for user interaction, which are also being addressed in F3.

Modern vehicles have a multitude of communication interfaces that serve as gateways for malicious hackers. For many years, our research has contributed to better protecting our vehicles and electronic road infrastructure against attacks, which requires continuous research and innovation in light of increasingly sophisticated attack techniques. Furthermore, smart vehicles are collecting and processing more and more data. This reinforces fears that data collection could ultimately lead to transparent drivers. Innovative approaches with data protection-friendly technologies can counteract this without significantly restricting the desired functionality.
More and more driving functions depend on the correct functioning of vehicle electronics and software and the exchange of correct data. Especially for data from external sources, it is imperative to ensure that the content is plausible, consistent, and trustworthy so that driving decisions can be based on it, for which appropriate evaluation mechanisms must be developed. In F3, these and other issues are being researched in order to make intelligent vehicles and traffic systems safer, more privacy-friendly, and more trustworthy.

The field of driver-vehicle interfaces investigates the interaction between humans and intelligent, automated vehicles. The focus is on how drivers can control their attention despite potential distractions, maintain reliable situational awareness, and receive targeted support from vehicle automation systems. Increasingly, the emphasis is shifting toward understanding drivers and vehicles as a team that shares responsibility. The design of takeovers, i.e., situations in which responsibility changes between humans and systems, plays an important role here. Interfaces should be designed in such a way that takeovers can be carried out efficiently or, in some cases, avoided altogether through cooperative approaches. Different forms of communication can help to strengthen trust, reduce mental stress, and support tasks. It is crucial that human-machine interaction is designed to be transparent and reliable. To ensure this, concepts are tested in simulations and studies and evaluated using objective criteria.

The field of human-environment-vehicle interaction research focuses on the interaction between automated vehicles and their social and physical environment. Particular focus is placed on communication with pedestrians, cyclists, and other road users, which is crucial for safety and acceptance. Road users must be able to rely on vehicles to react reliably even in risky situations, such as those involving children or visually impaired people. This trust also influences whether drivers are willing to initiate a takeover or trust the automation system. The assessment of criticality also depends heavily on the behavior of others, for example, whether pedestrians are distracted or attentive. Ambiguous situations pose a particular challenge in human-environment-vehicle interaction. Automated vehicles must communicate their intentions clearly and transparently so that people can adjust their behavior on the road without experiencing uncertainty or delays.