GreenBotAI: AI makes robotics efficient and sustainable

Franco-German AI funding is part of a long-term strategy that has been intensified since the Aachen Treaty of 2019. The German Federal Ministry for Economic Affairs and Energy and the French Ministry for Economic Affairs and Digital Sovereignty are jointly pursuing the goal of strengthening European competitiveness, promoting green and digital technologies and securing technological sovereignty. The GreenBotAI project, which makes robotic automation more efficient and robust, was also created within this framework. The four partners Fraunhofer IWU, Munich University of Applied Sciences, INBOLT SAS and ENSAM LISPEN presented the results of three years of intensive work at the German Embassy in Paris on September 24.

Crisis resilience through AI

GreenBotAI was selected in the "Innovation Projects on Artificial Intelligence Technologies for Risk Prevention, Crisis Management and Resilience" funding call. A total of around 17.9 million euros was invested in five projects aimed at crisis resilience through AI, particularly in the areas of sustainability and supply chains.

GreenBotAI: A lean AI model helps to significantly reduce energy consumption. © ENSAM/Arts et Metiers/LISPEN-Campus Lille

Against the backdrop of smaller batch sizes, more complex production lines, growing competitive pressure and unstable supply chains, GreenBotAI focused on faster response times, optimized path planning and the parallel execution of tasks during robot movement. The algorithms developed enable applications such as on-the-fly bin picking, assembly or quality control without the need for in-depth robotics knowledge. Modular machine learning models, trained in simulation environments with synthetic data, allow robust 2D and 3D tracking in combination with force-torque control. Thanks to UDP-based real-time communication and modular architecture, the solutions can be flexibly transferred to different robot systems.

Reduction in energy consumption

One focus was on energy efficiency: a reduction in energy consumption of over 25 percent was achieved through data-reduced AI models, accelerated gripping tasks and lower computing power - without replacing existing robots.

The consortium leader was Fraunhofer IWU with its expertise in process digitalization, production automation and AI. Munich University of Applied Sciences contributed its strong network with industry and research and teaches students interdisciplinary, practical work. INBOLT SAS, a Paris-based deep-tech start-up, contributed solutions for AI-supported 3D vision and robotic guidance systems, while the ENSAM research laboratory LISPEN contributed its expertise in the simulation and control of complex systems.