Festo teams for self-learning industrial AI robots
German robotics pioneer Festo has teamed up with Clevr in Amsterdam on self learning AI for industrial robots.
CLEVRAssist is a machine learning framework that allows machines to learn and adapt to their environments, offering a range of benefits to manufacturers, including increased productivity, reduced downtime and improved safety.
The tool is the result of a joint research project resulting from the collaboration with the Festo Learning Centre and the German Research Centre for Artificial Intelligence (DFKI). Together, the three partners have developed a prototype industrial assistance workstation that enables true interactive and situational learning. Festo designed the semi-automated assembly workstation with all the necessary automation components, while DFKI developed the sensor technology as well as the AI-based analysis of human movement data.
The prototype industrial assistance workstation provides instruction and guides users through each step of the partially automated process. Data generated in the work situation is collected by the automation components and the system recognises, for example, whether a certain part of the appropriate length has been installed, or whether a torque wrench has been used correctly.
This new generation of assistance and learning systems needs to be context-sensitive, intelligent, and adaptive to support workers to gain the more advanced qualifications they need, says Klaus Herrmann, Head of Research at the Festo Learning Centre.
DFKI has addressed the challenge of deploying artificial intelligence (AI) to identify and analyse movement patterns based on the sensor-based human movements captured at the assistance workstation.
DFKI developed an AI component that collects position data from sensor points on the wrists, elbows, and/or torso in a real-time activity tracking mode and performs process mining based on neural networks. The AI generates personalised recommendations for action that are tailored to the human user in the form of context-sensitive and personalised recommendations for the next work step.
The CLEVRAssist app brings together all the data flows and manages communication with the user. It operates at three main levels to achieve this: the cloud-based Mendix platform, IoT interface to collect all the data from the workstation’s peripheral devices, and digital mapping of data in its processes. The application handles communication and interaction with the learning human at the assistance workstation in accordance with didactic best practices and dynamically adapts the assistance to the personal needs of the user.
“We are excited to bring CLEVRAssist to market as the core of a new generation of digital assistance systems,” said Jeroen Hanekamp, CEO of Clevr. “With CLEVRAssist, we are leveraging the latest technology in low-code software and automation, combined with AI-based analysis of movement data, to provide a unique solution that enables true interactive and situational learning.”
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
