Research on Human-Robot Collaboration and Industry 4.0

Contact person: Dipl.-Ing. Dennis Mronga

Panorama view of the MRK Lab

Research Topics

Industrial dual-arm KUKA system
Industrial dual-arm KUKA system (Photo: DFKI GmbH, Annemarie Popp)
Example for human-robot collaboration
Example for human-robot collaboration (Photo: DFKI GmbH, Annemarie Popp)

In industrial manufacturing lines of the future as well as the area of consumer and retail, robots are working to an increasing degree side-by-side with humans and offer assistance. Thus, in the HRC Lab of the DFKI Robotics Innovation Center, we deal with research questions centered on Human-Robot Collaboration.

If robots operate in human environments without protective fences, the safety of the human has to be the highest priority. Because of this, we consider aspects on HRC-enabled robot control in the HRC Lab like active compliance or collision avoidance. The compliant robotic system COMPI for example is used to perform research on torque-based dynamic control approaches. The robot implements active compliance control by the means of its measured motor currents and a dynamic robot model. Another interesting control aspect is the area of mobile manipulation or Whole-Body Control, which makes use of all the robot’s degrees of freedom (e.g. mobile base and arm) in a unified way. This topic is currently investigated using Mobipick and the KUKA dual arm robot. Apart from control, we also investigate design aspects in the HRC Lab, like extreme lightweight construction or innovative motor concepts (e.g. passive-compliant actuation).

 Another safety aspect considers the “human-awareness” of a robot, which means recognizing the human and estimating its intention using sensors. In the Hybr-IT project for example, a person tracking algorithm has been developed based on data from 3D cameras and inertial sensors, worn at the human arm. By fusing this information streams the position and orientation of the human arm can be tracked in a robust way, which enables accurate handover of tools or materials from the robot to the human. The developed approaches have been implemented and tested on the dual-arm KUKA system in the HRC lab.

Experimental setup for the BEAR project
Experimental setup for the BEAR project (Photo: DFKI GmbH, Annemarie Popp)
Use of modern AI methods to learn „peg-in-hole“ task
Use of modern AI methods to learn „peg-in-hole“ task (Photo: DFKI GmbH, Annemarie Popp)

Apart from safety aspects HRC also deals with intuitive use and programming of robotics systems, which can be done by people without programming skills. Programming-by-demonstration, gesture and speech recognition, as well as augmented reality aspects have been investigated within different projects (e.g. TransFIT, HySociaTea) in the HRC Lab. In the iMRK project with the Volkswagen AG on the other hand, predictions on human movement speed and direction were used to adapt the behavior of the dual-arm KUKA system (e.g. its moving speed and compliance)

However, one of the most relevant current research topics  is the application of AI methods in industrial production. Under The buzzword “Industry 4.0”, this area considers the possibilities of digitalization for industrial manufacturing lines of the future. From the robotics point of view, the most interesting question is how robots can safely interact with the human in a highly interlinked, digitized production and benefit from the provided information to become safer, more flexible and autonomous. Projects like Hybr-iT or Knowledge4Retail illuminate these aspects.



last updated 31.01.2020
to top