How to help robots to understand the world
Human Robot Interaction research and activities within the group for Robotics and Semantic Systems focus on the coordination and communication of humans with robots and other (semi-) autonomous systems. We work mainly with industrial robotic systems, specifically those aimed at human robot collaboration such as ABB’s YuMi or Kuka’s iiwa, and marine vessels, such as unmanned surface vessels or submarines, but have also access to smaller mobile robot platforms and platforms targeting rather social aspects of interaction.
The driving force for our research is the assumption that we need to make robots understand their task rather than follow instructions blindly, and give them means to understand their own capacities and limitations in a given context.
We also want to support users of robots and other (semi-)autonomous systems in understanding these systems. This means in many cases that human users need to be supported in expressing what they want and mean when instructing a robot, as we could observe in various studies that instructions given to a robot can be incomplete and downright ambiguous, if not carefully interpreted in their context.
In short: we want to make robots understand that they do not understand and we want to support users in supporting robots and autonomous systems in accomplishing their tasks as best as possible.
Our approaches are based on classic AI methods such as knowledge representation and reasoning, probabilistic reasoning, as well as more recent approaches to combining these with learning approaches, both categorical and procedural.
If you want to learn more about HRI research at RSS, feel free to contact Elin A. Topp!
Individual projects and researchers connected to our HRI efforts
- WASP AS Industrial PhD Student Project A Digital Cognitive Companion for Marine Vessels
(Mårten Lager, Elin A. Topp, Jacek Malec)
- WASP AI PhD Student Project Robot Skill Learning Based on Interactively Acquired Knowledge Based Models
(Alexander Dürr, Elin A. Topp, Volker Krüger)
- Finished in 2017: SARAFun / Prace related PhD project Intuitive Instruction of Industrial Robots - A Knowledge Based Approach (Maj Stenmark, Elin A. Topp, Jacek Malec)