Current factories come from times of large quantities and programming, in which the robot can perform a specific task perfectly and as often as required. However, customer requirements have changed and are becoming increasingly small-scale. The factory will therefore also change. "In order to be able to design and quickly manufacture specialized, individual and often mechatronic products at reasonable costs, we need a scalable, flexible factory and robots that are capable of learning," says Prof. Klaus Bengler from the Technical University of Munich.

Creating this factory of the future is the core objective of the KI.Fabrik lighthouse project launched in mid-2021, which was initiated by the Executive Director of the Munich Institute of Robotics and Machine Intelligence (MIRMI), Prof. Sami Haddadin, and whose research and development projects are headed up by ergonomics professor Bengler. This new approach is now being put into practice in the new research laboratory.

The technological basis is provided by digital twins and robotic hardware controlled by an AI platform. Bengler wants to use this system to show that robots can carry out tasks independently, are capable of learning and can assist humans. "It's the human expertise that counts," says psychologist Bengler, "the robot acts as a teammate." In MIRMI's new robotics lab, almost 100 robotic arms are currently busy passing cables to each other, copying what scientists show them, learning together and assembling gear parts.

Scientist Keija Chen observes how her robot moves cables into the correct position.

© Andreas Heddergott / TUM

The doctoral students and post-docs from eight TUM chairs are focusing on eight so-called cases. From teleoperation and collective learning to network design, important scientific topics are being researched. And the capabilities of the adaptive robots are constantly growing: The robot avoids collisions, communicates over long distances via teleoperation or cooperates with other robots. All capabilities are also available to other robot systems via a software platform, regardless of the platform. The robot can also apply certain skills to comparable tasks. "The robot trains independently and transfers its knowledge," says Prof. Bengler.

From the industrial side, the car manufacturer BMW, the drive technology manufacturer Wittenstein, the robotics companies Franka Emika and Reactive Robotics, the electronics company TQ-Systems and the mechanical engineering company Linde Material Handling are supporting the technical developments in the KI.Fabrik lighthouse project. TUM is represented in the consortium by the Departments of Mechanical Engineering, Electrical Engineering and Information Technology and Computer Science, among others. A total of around 13 million euros is available for the overall project until 2025, which will primarily be invested in the development of infrastructure. A total duration until 2030 is planned.