The Bremen Institute of Structural Mechanics and Production Systems (bime) has designed and implemented a learning factory for human-robot collaboration (HRC), which enables the investigation of collaborative assembly processes within a clocked flow production.

Assembly line production is characterized by structured processes for which there is little knowledge about how collaborative robots can be optimally integrated into an existing cycle. This is where the bime learning factory comes in, not only to provide employees with practical training, but also to gain a deeper understanding of this specific form of collaboration and to optimize it. The basis of the learning factory is an experimental assembly system in which manual and collaborative workstations can be linked in a modular fashion. A special feature of the assembly scenario introduced is that a robot is not tied to a fixed assembly workstation. Instead, it can flexibly support activities at two manual stations. The technical implementation is carried out using image recognition to analyze the work progress and existing buffers at the individual assembly stations. In various operating modes, employees can either decide for themselves when they need support in the assembly process or rely on the system to prioritize their tasks.

Learning factory as a living laboratory

The learning factory functions not only as a training environment, but also as a living laboratory for scientific research. A central focus is on the influence of different degrees of collaboration on the cooperation between humans and robots. This investigation sheds light on questions relating to increasing efficiency and simultaneously ensuring safety in collaborative working environments. Another important scientific focus is on the arranging behavior within human-robot teams. The Learning Factory makes it possible to investigate the interactions between human employees and collaborative robots in an industrial environment. The focus of the analysis is on communication mechanisms, task distribution and decision-making in real time. The learning factory enables the transfer of knowledge from research to industry on several levels. Firstly, the trained employees can integrate their acquired knowledge and experience directly into their day-to-day work. This helps to ensure that the implementation of human-robot collaboration is smoother in practice and that employees are better prepared for the challenges of this modern working environment.

Innovation catalyst for research

On the other hand, the research platform of the Learning Factory enables continuous improvement of human-robot collaboration itself. By continuously collecting data and experience from the real production environment, existing models and concepts can be optimized and adapted to the specific needs of the industry. This cyclical process of optimization ensures that the knowledge gained in the learning factory is not only state of the art, but can also drive future developments. The Learning Factory thus acts as a catalyst for innovation, driving not only training but also the continuous further development of human-robot collaboration in research and industry.

In addition to the physical learning factory, there is also a digital twin of the assembly plant, the aim of which is to enable a virtual learning experience. Technologies from game development enable real assembly processes of the learning factory to be tested cooperatively with other learners in training scenarios in a virtual reality environment. The special feature of the solution developed at bime is that the software is largely independent of the user hardware. By using render streaming technology, content can be calculated on a central server and played back on any end device such as tablets or smartphones. This paves the way for location-independent education and training - whether in the office or on the sofa at home.

This article appeared in issue 12/23