Robots and automated systems are increasingly shaping the economy, society and everyday life. Whether in production or logistics, in mobility, medicine or at home: as intelligent assistance systems, they can support humans, take over certain routine activities completely or master previously unrealizable tasks. If robots are to act autonomously, there is no way around artificial intelligence (AI). At the technology day “Interactive, self-learning, autonomous - research at the KIT for the robots of tomorrow” of the KIT Business Club on May 14, 2025 at the Research Factory on KIT Campus East, guests learned how close we already are to the future with learning, autonomous robots and what is being worked on at full speed at KIT. The event focused on five KIT researchers and their research work, which showed how intelligent and versatile robotics can already be today.

The path from AI-created cooking instructions to autonomous cooking robots is a long one and will still require many development steps. This is how Prof. Katja Mombaur from the Institute for Anthropomatics and Robotics (IAR) described the current state of the art at the start of the event. Nevertheless, AI methods are currently making great progress in key robot capabilities such as perception, motion control and planning. Various applications in the KIT BioRobotics Lab proved that their use is worthwhile: exoskeletons that make walking easier for people with restricted mobility, humanoid “social robots” as assistants or a bouncing robot dog as a companion and navigator.

In view of the ever-increasing variety of products and changing requirements, such as the circular economy, the machines of the future in production and logistics processes must be able to handle varying objects flexibly and, if possible, without new programming or training. With her approaches to using AI to teach robots, Prof. Rania Rayyes from the Institute for Materials Handling and Logistics (IFL) is giving robotic arms greater adaptability and more autonomy, for example when it comes to gripping a wide variety of objects or finding them.

Is it possible to reconstruct the original photo motif from a completely noisy image? And why can this advance robotics? Prof. Rudolf Lioutikov (Intuitive Robotics Lab at the IAR) demonstrated this using “reverse diffusion models”. His research group uses these AI models to program robot movements, combines them with large language models and showed that they can be used flexibly and require less computing power than other approaches. If this finding is confirmed, the method would also be particularly suitable for decentralized systems and for small and medium-sized companies.

Particularly impressive for the guests: in the laboratories for Machine Intelligence and Robotics of Prof. Arne Rönnau and Decision Making and Data Science of Prof. Anne Meyer of the Institute for Information Management in Engineering (IMI), they were able to experience at first hand how autonomous walking robots and driverless transport systems work - technologies that already move and make decisions amazingly independently today. AI also plays a key role in the development of a new generation of these machines for sensor technology, control and, above all, decision-making. However, the newly acquired intelligence is closely linked to even more powerful hardware, such as batteries and motors. These are essential if such systems are to be used for autonomous delivery processes or inspections in the future.

If a single robot is not sufficiently equipped for a task, a team of robots may be able to help: Participants were able to discover impressive examples of this at the technology day. For example, IMI successfully took part in an ESA competition to explore unknown environments with a trio of walking and driving autonomous robots. A team of cooperating robots can also be helpful for industrial applications. Malte Hansjosten from the wbk Institute of Production Science demonstrated this using the example of dismantling battery packs. And the KIT spin-off Formic, whose swarm robots can transport heavy loads through tight spaces together, made it clear that they are stronger together than alone. The founding teams Prefiro with harvesting robots for special crops and the plant robots of the student university group Kamaro Engineering e.V. also showed how robot technologies can withstand physically demanding conditions.

The time for sophisticated assistants with metallic stature, code in their hearts and electronic drive therefore seems to have arrived. In order to take the next groundbreaking steps and keep up with the rapid pace of international development, Prof. Rudolf Lioutikov and the other researchers from the KIT in particular would like to see a stronger alliance and more intensive cooperation with industry in Germany and Europe. The first opportunity for the researchers and the industry guests to connect was during the final networking session, where they had a lively exchange about their impressions of the technology day and made new contacts. The successful conclusion perfectly matched the goal of the KIT Business Club: to connect people from science and industry. After all, despite all the technology, in the end it is people who start cooperative projects. This is aptly summed up by the following quote from Commander Data from the Star Trek series: “I am superior, sir, in many ways, but I would gladly give it up to be human.” (Star Trek: The next generation, Encounter at Farpoint, Paramount Pictures, USA, 1987).