by Andreas Hemmerle, Fraunhofer-Institut für Werkzeugmaschinen und Umformtechnik IWU
Smaller batch sizes as an alternative of mass manufacturing, extra advanced manufacturing strains, rising aggressive stress, unstable provide chains: Towards this background, the Franco-German analysis venture GreenBotAI addresses robotics.
Three important targets are in focus—guaranteeing manufacturing continuity even in pandemic occasions, strengthening Europe’s independence in manufacturing automation, and considerably lowering the vitality consumption of robotic functions in European factories.
Technically, GreenBotAI focuses on the response and latency occasions in industrial robotics, optimized path planning, and executing well-defined duties whereas the robotic is in movement (“on the fly”). With out stopping, for instance, it’s attainable to seize 2D photographs for object recognition for the specified dealing with or meeting activity. On this Franco-German analysis venture, Fraunhofer IWU leads the consortium.
The venture focuses on an agile method to dealing with supplies and workpieces. AI algorithms are supposed to present manufacturing extra flexibility and speed up manufacturing in numerous industrial sectors. The purpose is clever robotics able to dealing with advanced duties spontaneously.
Required elements in {hardware}, state-of-the-art deep studying strategies for monitoring, enhancements in knowledge processing and error management, are anticipated to herald a generational shift in robotics.
GreenBotAI is engaged on a number of levers to cut back vitality consumption by as much as 25%. These embrace data-reduced AI fashions, accelerated gripping duties, and lowered computing energy.
On the Hannover Messe, held April 22–26, 2024, the venture companions will reveal what they’ve achieved as but. A person picture taken in 2D with industrial digicam know-how (“One-Shot Demonstration”) of the part and little computing energy are enough for the robotic to select up, examine, and place this part as desired.
The software program Xeidana, developed at Fraunhofer IWU, verifies whether or not the robotic has picked up the right part. Company on the sales space can attempt it out themselves: They place a gear at any level on a desk within the workspace of a collaborative robotic (a robotic designed for collaboration with folks).
Primarily based on the picture data, the software program determines the positions of all objects relative to the robotic, calculates the robotic’s path, and determines the gripping place. The robotic then picks up the gear, and inspects it, relying fully on the picture data.
That is the place real-time analysis of drive knowledge comes into play, guiding the robotic on how one can match the picked-up gear right into a second gear, utilizing a wise software of AI that mimics human contact. A digital twin moreover visualizes all actual actions of the robotic.
Guests may also expertise the meeting and built-in high quality management software reside.
Offered by
Fraunhofer-Institut für Werkzeugmaschinen und Umformtechnik IWU
