"Our biologically-inspired approach involves modeling the neural connections in the spinal cord, found in simple vertebrates like lampreys or even in humans, to establish the fundamental movement pattern," explained Hausdörfer. "Additionally, we model local feedback, which is learned using reinforcement learning. This approach was implemented on a snake robot consisting of ten segments." Through this implementation, robots can also provide insights into biological neural networks in animals.

Reinforcement Learning: Learning from Lampreys

Hausdörfer and his team carried out his research at TUM and the Swiss Federal Institute of Technology Lausanne (ETHL). They showed that the system learns significantly more robustly and finds better solutions based on this approach compared to traditional methods. For the Young Scientist prize, the winner was determined through votes from the audience on-site and online. Dr. Tina Klüwer, head of the Department of Research for Technological Sovereignty and Innovations at the BMBF, presented Hausdörfer with the award at KIRO 2024.