Developing image data-based AI algorithms. KARLI & Fraunhofer IOSB

Within KARLI, the Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB) is developing image-based AI algorithms. These are designed to detect when drivers and other occupants are behaving as they should in automated vehicles by the current level of automation.

Author: Dr. Frederik Diederichs

With 30,000 scientists at more than 70 institutes in Germany, Fraunhofer is Europe’s largest research organization. We work primarily on the transfer of research results into applications. By doing so, we always try to conduct research as close as possible to the current needs the industry. We develop organizational and technical solutions for our customers’ current and future challenges. As we have a good understanding of the market, we pursue strategic goals for applications that will hopefully become relevant in the future.

Most of our work is in publicly funded consortia projects, similar to KARLI. In this case, we even initiated the project ourselves, together with solid partners who want to reposition or improve their applications and processes with the help of artificial intelligence.

In this case, we recognized the need for new approaches to the intelligent vehicle interiors of the future, defined them, and found partners who share this roadmap to shape it with their own goals. Within the consortium, we work closely together, exchange ideas, learn from each other, and are thus able to advance the topic more efficiently. The project is also supported by the German Federal Ministry for Economic Affairs and Climate Action (BMWK).

In KARLI, the Fraunhofer Institutes for Industrial Engineering (IAO) and for Optronics, System Technologies and Image Exploitation (IOSB) are involved.

The “Interactive Analysis and Diagnosis” (IAD) department at Fraunhofer IOSB develops innovative interaction methods and camera-based assistance systems for vehicles. For example, IOSB’s Advanced Occupant Monitoring Systems development is based on image analysis for person perception for adaptive human-machine interactions (Perceptual User Interfaces). Its own Deep Learning AI methods enable the classification of activities and intentions for intelligent and proactive vehicle interiors and interaction with smart assistants.

Within KARLI, IOSB is developing image-based AI algorithms to detect level-compliant behavior and determine motion sickness in vehicle occupants. Among other things, detected activities and intentions improve interactions with vehicle occupants.-being at an early stage.

For Fraunhofer IOSB, one of the goals in KARLI is to generate an excellent and value-stable data basis for our AI algorithms. We work closely with our colleagues at Fraunhofer IAO and with Continental, Ford, and Audi, who put KARLI vehicles on the road.

About the author

Dr. Frederik Diederichs is a Senior Scientist at Fraunhofer IOSB responsible for the strategic initiation of research and implementation projects focusing on AI image analysis and driver state recognition.

In the KARLI project, as a technical-scientific coordinator, he is primarily responsible for the content-related orientation and good working conditions for the partners involved. Together with consortium leader Christoph Wannemacher from Continental, he assists the partners with organizational issues and supports cooperation and exchange.

He is driven by research into safer, more comfortable mobility and how to make the most of travel times. Interacting with technology should be fun.

Leave a comment