In collaboration with EDAG Production Solutions, the Chair of Production Systems at Ruhr University Bochum designed and built a fully networked, hybrid assembly system including a functional virtual image. The focus of the research project is on the optimization of individual subsystems and the simulative design of workstations for human-robot collaboration.

The COssembly hybrid assembly system consists of a total of eight industrial robots and various assistance systems. Five of the eight robots are fully designed for human-robot collaboration (HRC), which allows them to work together at workstations without safety fences. In view of the research focus on hybrid assembly, two manual workstations are also integrated into the production system. On this basis, generator connection boxes for photovoltaic systems can be manufactured as typical products of the electronics industry in an HRC environment.

The virtual image of the assembly system

COssembly stands out from industrial production facilities in particular due to its heterogeneous system structure. The focus is on the interaction between mainly HRC-capable robotics from various manufacturers and powerful assistance systems. For example, the employee works directly with a Fanuc CR-7iA to fasten cable glands, while the Kuka KMR iiwa mobile robot system pre-positions the electrical connection elements for manual assembly before ABB's YuMi collaborative two-arm robot performs the final voltage test. However, the scope of the project goes far beyond the construction of a functional demonstrator. With the aim of being able to master the virtual commissioning of such a heterogeneous overall system and thus address a problem of practical relevance, the project partners created a functional virtual image of the COssembly. As this exactly replicates both the geometric characteristics and the behavior of the real system, the safety concept and the functionality of the interlinked subsystems could be simulatively validated in advance.

Current and future research projects

In view of the variety of robot and assistance systems used and the modular design of the assembly stations, the COssembly hybrid assembly system is predestined for investigating various issues from the fields of industrial and basic research. Among other things, the potential of the flexible use of the mobile robot platform is currently being evaluated. In addition to the development of strategies for automated material provision, the fine localization of the Kuka KMR iiwa mobile robot in the production environment, which is essential for this, is currently the subject of research work. Furthermore, the range of applications for the robot platform is to be extended in future to include mobile maintenance and servicing of the hybrid assembly system. For example, if information is available about a faulty orientation of the generator connection box, the mobile robot system should navigate autonomously to the relevant station and rectify the fault independently. If such status data is finally fed back into the virtual image, the result is the digital plant twin, which maps the process behaviour of the real plant in real time and thus enables the comprehensive optimization of the real process.

Although there is a complete download option between the COssembly and its virtual image, there is currently no bidirectional data exchange. In order to be able to enrich the virtual image with real-time sensor data from the real system and thus update it, additional sensor technology needs to be integrated. Furthermore, the virtual image will be used in the future to develop planning tools for hybrid assembly systems with HRC workstations. In this context, both the design of workstation environments for safe task switching between humans and cobots and the simulation of new HRC safety concepts will be considered. These findings will ultimately be incorporated into a training program and thus contribute to a further increase in acceptance with regard to the use of collaborative robots in workspaces without safety fences.

P. Kulessa, M. Boshoff, B. Kuhlenkötter, LPS, Ruhr University Bochum