In order to minimize the limitations in human-robot collaboration and meet safety requirements, the Fraunhofer Institute for Cognitive Systems IKS is conducting research on “Adaptive safety”. This refers to methods and algorithms that - in connection with sensors and actuators - enable adaptable safety. Instead of static safety functions, adaptive safety functions are to be used in the future.
For example: A static safety function follows the principle “When the safety door is opened, then stop the system.” With an adaptive safety function, on the other hand, it looks like this: “When the safety door is opened, then switch to reduced operation for the immediate proximity. Operate all other areas normally.”
For different application areas, there are different scenarios for how Adaptive Safety can be implemented. A collaborating robot, also called a cobot, can detect whether an obstacle is in the way using torques or energy consumption measurements at the axle motors. Cameras or an artificial skin also allow the system to detect when it is touching a person, for example. However, Adaptive Safety goes one step further: Intelligent software enables the robot to assess the behavior of humans in advance and thus plan its movements with regard to the necessary safety requirements. For example, it would be possible to optimize the route or adjust the speed, even in case of driverless transport systems.
Another critical point in production is the intervention in productive operation, for example when a sensor must be readjusted but the current batch has to be further processed. Thanks to Adaptive Safety from Fraunhofer IKS, the machine does not have to stop the entire production. Thanks to sensor-based monitoring, it only secures the area in which people work. If the human leaves this zone again, the machine automatically switches to normal operation.
The vision of Fraunhofer IKS are cognitive production facilities and factory halls: In a cognitive factory hall, all objects, people and machines, can be localized and their intentions are recognizable. Humans and robots can thus work efficiently, safely and collaboratively.