MRK
Robots recognize gestures and facial expressions
The Fraunhofer Institute for Machine Tools and Forming Technology has developed a technology that makes working with large robots more intuitive. It enables robots to recognize gestures, faces and body postures.
If you take a look at production halls, collaboration between humans and robots is commonplace. Or rather: working side by side. Although heavy-duty robots now work alongside their human colleagues without a safety fence, direct interaction is not possible. For safety reasons, the robot stops moving as soon as the human enters a large safety area around it.
Safe, effective and direct collaboration
A new type of technology from the Fraunhofer Institute for Machine Tools and Forming Technology IWU is now increasing the efficiency of this collaboration - and therefore of the entire production process. "We have added effective, safe and flexible interaction to the technology used to date," says Dr. Mohamad Bdiwi, Head of Department at Fraunhofer IWU. "For the first time, humans can communicate and cooperate directly with heavy-duty robots."
In production, this cooperation looks like this: When a person enters the work area around the robot, the robot recognizes their gestures, face and posture. The data is used to ensure safe collaboration and also for control purposes. For example, humans can give their metal colleague work orders using hand and arm gestures - the robot analyzes even complex movements. "Our technology brings gesture control to the industrial sector. Until now, it has primarily been used in gaming environments, such as consoles," adds Bdiwi.
In addition to the human's hands, the robot also keeps an eye on their face: if the human looks to the side or back because they are talking to a colleague standing there, for example, the machine knows that the arm movements are not directed at them. Humans and robots can work together directly and also pass workpieces or tools to each other. If the worker's hand is too close to his or her face, so that a handover would be dangerous, the robot recognizes this and waits until the worker extends his or her hand far enough. At the heart of this human-robot interaction are intelligent algorithms and 3D cameras that serve as the robot's "eyes". as










