A key research topic is collision avoidance and path planning for dynamic testing and manufacturing tasks. Robot motion planning is specifically tailored to the object at hand and optimized with respect to constraints (reachability, pose-dependent loads) to shorten cycle time or enable it.

For robot-based component testing, force and torque data are actively used for control. This allows for the adjustment of unwanted load components or active trajectory tracking to maintain a specific load vector.

In the area of robot-based inspection, three collaborative robots can perform various testing methods (including radiography, computed tomography, optical inspection) on large objects (e.g. complete vehicles). The focus is on optimizing the testing process and data fusion from different measurement systems.

For robot-based bending, force- and torque-controlled dual-arm systems of various sizes are used. The kinematic flexibility of the collaborative robots allows for manufacturing significantly more complex structures without additional joining processes. Real-time data from force-torque sensors are used to optimize the process with artificial intelligence models.

Similarly, various sensor systems are used in the field of robot-based friction stir welding to monitor and optimize the process in real time with the help of artificial intelligence. The goal is to optimize trajectories, reduce rejection rates, and enable quality assurance from process data without additional testing effort.

Another focus is on transferring simulation models into real robot applications ("Sim2Real"). Digital models are used for planning, optimization, and virtual commissioning, with a special emphasis on accurately representing sensor data and environmental conditions. Validation and feedback of real measurement data into the simulation environment improve both model quality and process safety.