Contact Person: Dr.-Ing. Dennis Mronga
In future, robots will increasingly work side by side with humans and assist them in industrial manufacturing as well as in service robotics. In the HRC laboratory of the Robotics Innovation Center, research questions related to physical human-robot interaction and collaboration are addressed.
If robot systems are operated without separating protective devices, the safety of humans in the vicinity must be ensured. Therefore, various safety aspects of collaborative robotics are researched in the HRC laboratory. This includes methods for safe robot control such as active compliance and collision avoidance. The COMPI robot system, for example, is used to research torque-based dynamic control methods. It implements compliance control based on the measured motor currents and using a dynamic model of the arm. Another interesting aspect is mobile manipulation, i.e., the joint consideration of the degrees of freedom of the mobile base and the robot arm, which is being researched on the Mobipick robot, for example. In both cases, fundamental insights from modern control theory are brought into practical application, for example from the areas of dynamic/model predictive control or whole-body control. In addition to the implementation of control engineering methods, the HRC laboratory also considers design issues, such as extreme lightweight construction and innovative motor concepts (e.g., quasi-direct drives and passive compliant actuators). Another safety aspect concerns the question of "human awareness" of robots, i.e., the recognition of humans and their intentions via suitable sensor technology. In the Hybr-iT joint project, for example, a person tracking system was developed based on 3D cameras and inertial sensors worn on the human body. In this way, the position and orientation of the human arm can be detected, which enables a precisely fitting transfer of tools or components between humans and robots. The method was implemented and tested on the iMRK robot system.
In addition to safety aspects, HRC also raises the issue of intuitive use and programming of robots, which can be done by humans without programming knowledge. Imitation learning and robot control based on gestures, as well as augmented reality aspects have been researched for this purpose in various projects (e.g., TransFIT) in the HRC lab. In the iMRK project with Volkswagen AG, the robot system of the same name was used to make predictions about human behavior for robot control. For example, the speed and direction of movement detected by laser scanners was used to actively adjust the speed and compliance of the system.
Another relevant research topic is the application of AI methods in production. Summarized under the buzzword "Industry 4.0", this area looks at the possibilities of digitalization for industrial manufacturing of the future. This raises the question of how robots can interact with humans in a fully networked manufacturing environment and benefit from the information provided to become safer, more flexible, and autonomous. Projects such as Hybr-iT or Knowledge4Retail are shedding light on these aspects.