Universität Bonn

Institute of Computer Science

22. April 2024

Service robots of the NimbRo team win German Open Service robots of the NimbRo team win German Open

At the RoboCup German Open Championships in Kassel, the service robots of the NimbRo team from the University of Bonn achieved the highest score in the @Home league.
At the RoboCup German Open Championships in Kassel, the service robots of the NimbRo team from the University of Bonn achieved the highest score in the @Home league. © Autonomous Intelligent Systems/University of Bonn
Download all images in original size The impression in connection with the service is free, while the image specified author is mentioned.

At the RoboCup German Open 2024 from April 18 to 20 in Kassel, the service robots of the NimbRo team from the University of Bonn achieved the highest score in the @Home league during the conducted tests and also convinced the jury in the final. The service robots were developed in the Autonomous Intelligent Systems research group at the Institute of Computer Science, supported by the Lamarr Institute for Machine Learning and Artificial Intelligence. They navigate autonomously, can grasp and deposit objects, and interact with people through a speech-dialogue system.

The RoboCup German Open is the largest annual European competition for intelligent autonomous service robots in the @Home league. The goal is to develop assistant robots for everyday environments, such as assisting people in need. The robots' abilities are assessed through a series of tests in a realistic home environment. Emphasis is placed on intuitive human-robot interaction through speech and body language, swift and safe navigation in unstructured everyday environments, and object handling.

The tasks tested are diverse: Robots should, for example, greet guests and enforce house rules, store purchases and set a breakfast table, as well as serve guests in a restaurant and clean up the kitchen. The Bonn NimbRo team, under the leadership of Prof. Dr. Sven Behnke, programmed a mobile robot with a flexible chassis and two arms for these tasks. It perceives its environment through cameras, laser scanners, and a microphone, and controls its 23 motors and a loudspeaker to independently perform the assistance tasks. Methods from AI research are used, such as image understanding, speech understanding, motion planning, and dialogue systems.

The Bonn team took the lead right from the first test and maintained it throughout the three-day competition. In the final, the two best teams could demonstrate a self-defined task. Here, the Bonn robot demonstrated how it navigates in the apartment, recognizes ingredients, language, and gestures, and supports the user in preparing dinner based on the recognized ingredients through recipe ideas. This convinced the jury, composed of selected experts and other team leaders.

"The development of assistant robots is important to enable people in need to live independently in their familiar environment," says Prof. Dr. Sven Behnke, head of the Autonomous Intelligent Systems research group at the University of Bonn and Area Chair for Embodied AI at the Lamarr Institute. "The competition in Kassel provided important insights for our further research," says the scientist, who is also a member of the Transdisciplinary Research Areas "Modelling" and "Sustainable Futures", the PhenoRob Excellence Cluster, and the Center for Robotics.

Prof. Dr. Sven Behnke
University of Bonn
Institute of Computer Science VI – Autonomous Intelligent Systems
Phone 0228 / 73 4116
Mail: behnke@cs.uni-bonn.de

Wird geladen