Développement de robots d’élagage adaptatifs pour une viticulture durable grâce à la simulation et à la perception assistée par l’IA, à l’apprentissage robotique et à la téléopération intuitive – DEMETER

The DEMETER project aims to advance viticulture by integrating cutting-edge technologies in robotics, computer vision, machine learning, simulation, and teleoperation to create a highly advanced, adaptable vision-guided pruning robot. This project addresses the critical need for efficient and sustainable agriculture practices by developing a system that can perform vine pruning autonomously and accurately, minimizing the environmental impact and reducing dependency on manual labor. By leveraging a dual workflow that includes both offline preparation using detailed 3D simulations and online execution with real-time data calibration to fit the simulated vineyard environment with the input real data, DEMETER aims to enhance the precision of vine pruning operations under various environmental conditions. The integration of state-of-the-art perception systems and feedback control mechanisms ensures that the robotic operations adapt seamlessly to real-world variability, setting new standards for automation in viticulture and demonstrating significant advances over existing technologies. This integrated approach promises not only to increase productivity but also to foster the transfer of skilled pruning techniques from humans to machines, ultimately redefining the landscape of agricultural robotics. In gathering world leading experts and key stakeholders from Europe (France and Germany) and Japan with their complementary expertise, the DEMETER project is fully aligned with the goals of EIG CONCERT-Japan in fostering collaboration and transnational mobility between European countries and Japan, and enhancing cooperation in science and technology fields through this multilateral joint funding initiative.