Safety in Industrial Robotics Fraunhofer IFF Demonstrates Adaptive Safety Logic for Industrial Robots

Related Vendors

Fraunhofer-Institut für Fabrikbetrieb und -automatisierung IFF

FAULHABER Drive Systems

Wuxi InfiMotion Technology Co., Ltd.

In the future, robots should react flexibly to changes, work together with humans and remain predictable. A new case study by the Fraunhofer IFF shows how this goal is being technically implemented in the EU project RoboSAPIENS and what this means for industry.

However, industrial robots quickly reach their limits when their environment changes. New product variants, unplanned malfunctions or direct collaboration with humans have so far required costly retrofitting or additional safety zones. This is precisely where the European research project RoboSAPIENS comes in, which has been running since January 2024 as part of Horizon Europe.

The aim of the project is to enable robots to adapt their strategies themselves without sacrificing safety and traceability. Unlike traditional systems, the RoboSAPIENS approaches do not work rigidly, but use AI methods and virtual models to evaluate situations and adapt their behavior. The digital twin plays a key role here, first running through changes virtually before they take effect in the real system.

Practical Testing in Magdeburg, Germany

The Fraunhofer Institute for Factory Operation and Automation IFF in Magdeburg, as a partner in the field, is putting the concepts developed into realistic application scenarios. The focus is on human-robot collaboration with cobots in flexible production environments, i.e. where traditional safety concepts often reach their limits.

The recently published case study shows that robots can adapt their behavior during ongoing processes, for example in the event of disruptions or unexpected changes in the workspace. Digital twins serve as a safety net: adaptations are simulated in advance and risks can be identified and avoided at an early stage. In addition, special mechanisms ensure that human safety is also guaranteed during the adaptation process.

Dynamic Risk Asessment Instead of Blanket Speed Reduction

In technical terms, the project combines the "force and power limitation" and "speed and distance monitoring" safety modes in accordance with ISO/TS 15066. In practice, this means that the robot operates at high speed if there is no human nearby. If a person approaches, the system reduces the speed depending on the part of the body, but always remains within the permissible limits. The aim is productive collaboration without any overall loss of speed.

A camera system captures the human in the workspace as a three-dimensional skeleton and differentiates between areas of the body such as the hand, upper arm or head. The system links this position data with the robot's current movement and position information and uses it to continuously calculate the risk. The robot adjusts its speed depending on the situation: When close to the hand, it reduces its speed to such an extent that a potential collision does not become painful; when approaching the head, a significantly stronger throttling takes effect. In addition, the approach distinguishes between impact and clamping by taking into account whether a clamping point with a static structure can occur at all. The robot should therefore work quickly without a protective fence, but react in a controlled and comprehensible manner when close to people.

Step Towards Industrial Use

For industrial companies, this could be an important building block for further developing automation and human-robot collaboration without having to rely on rigid safety concepts. Whether and how quickly the concepts will find their way into series solutions should become clear in the coming months.

RoboSAPIENS is a 36-month EU research project. Coordinated by the University of Aarhus, universities, research institutions and industrial partners from several countries are working together. In addition to the Fraunhofer IFF, the University of York, the University of Antwerp, the Aristotle University of Thessaloniki, the Danish Technological Institute and Pal Robotics are also involved. (mc)

Secure and Compliant Authentication in Laboratories

Secure & Contactless User Authentication at GMP/GLP