In the AI Innovation Center “Learning Systems and Cognitive Robotics”, IPA researchers are developing the “Universal Safety Framework.” This aims to address current safety challenges such as rising costs, lack of flexibility, inadequate performance and insufficiently transparent AI technologies and also to provide new solutions.
The basis for this is a safety toolchain, which comprises planning and analysis tools – primarily for human-robot collaboration applications – as well as tools for commissioning and operation. These include IPA developments such as CARA, the F-PLC Configurator and the HRC test bench. In the future, these tools will be networked with each other.
A digital safety shadow complements this toolchain so that routes for automated guided vehicles can be optimized, risk analyses evaluated, or the safety configurations of stationary robots dynamically adapted, etc. To do this, data relevant to safety such as laser scanner frequencies, causes of collisions or robot speeds are recorded and analyzed in the safety shadow.
Additionally, AI is used to advance from detection to “human perception.” Approaches to validate AI for safety-critical applications are being explored. Based on the generated and analyzed data, a safety management process takes place that feeds back outputs and commands to the toolchain.
If you would like to find out more about the Safety Framework or are interested in being involved in its development, please contact the project manager Thomas Koch by phone or by e-mail as shown in the right-hand column.