Skip to main content

Detail view

How Does AI Drive a Car Through Chaotic City Centers?

Together with two other universities and the technology company Continental, the University of Bremen is carrying out research on automated vehicles. The PRORETA project deals with the recognition of complex traffic situations in city centers.

The name PRORETA was chosen as a nod to the eponymous crew of Roman war ships. The proreta stood in the bow of the ship (Prora) and warned of shoals and other hazards. Alongside the University of Bremen, the Technical University of Darmstadt is involved as project coordinator for the fifth research round, as is the Gheorghe Asachi Technical University of Iasi, Romania. The demonstration vehicle for the tests is being constructed and equipped by Continental.

Challenge: Uncontrolled Crossings

The aim of the fifth phase of the PRORETA research project, which is to last until 2022, is the development of algorithms. They are to deduce correct driving decisions from the sensor data that are comparable to human decisions. For example, it is a challenge to interpret all objects relevant for the driving direction at an uncontrolled crossing. This includes their direction of movement, intentions, and priority in traffic. Without human aid, artificial intelligence (AI) is to make confident decisions. The major advantage of AI is that after a training phase, it is in the position to make the right decisions in unfamiliar situations based on what it has learnt. One of the sub sections of the project will be the observation of human drivers when they are reducing and assessing the complexity of surroundings. The algorithms that are capable of learning are to be trained along similar principles.

Recognition of Surroundings as Special Field of Expertise

At the University of Bremen, the Cognitive Neuroinformatics working group, which is under the direction of Professor Kerstin Schill, is involved. The group focusses on elementary cognitive abilities within the PRORETA project, for example self-localization, object recognition, and object following. The working group therefore has special expertise in the field of surroundings recognition through sensor data fusion. Those are all processes that are connected to awareness and recognition. “Transferring human cognitive abilities onto the “intelligence” of vehicles so that they can better deal with complex situations is the big challenge of this project,” says Professor Kerstin Scholl. Currently, the first measuring campaigns to record training data in the prototype vehicle are taking place in Bremen.


Prof. Dr. Kerstin Schill
Faculty of Mathematics/Computer Science
Cognitive Neuroinformatics Working Group
Phone: +49 421 218 64240




Drawing of a car at a blind intersection
AI is to interpret relevant objects in city center traffic with the help of algorithms.