Imagine a robot that learns how to do a new task just by watching a video, rather than needing a programmer to write thousands of lines of code. It's like giving a factory arm a pair of smart eyes and a brain that can handle objects it has never seen before. This allows a robot to switch from making a satellite part to a custom package almost instantly.
Industrial robots are currently too rigid, floor-fixed, and slow to reprogram for small-batch production. This leads to costly downtime and material waste in sectors like aircraft and satellite manufacturing.
An open-source AI system for mobile robots featuring a task planner that learns from video and a perception system for unseen objects.
If you are a satellite manufacturer dealing with high-mix production where every part is different — this project developed an AI-powered system that enables a 15% productivity increase and 20% less material waste.
If you are a packaging company dealing with frequent changes in product variants — this project developed a versatile robotics system that reduces downtime by allowing robots to adapt to unseen objects.
If you are a lift manufacturer dealing with rigid, floor-fixed robots that are slow to reprogram — this project developed a mobile manipulator system that increases battery autonomy by 5%.
Based on available project data, specific pricing or implementation costs are not provided.
The project is designed for industrial scale, featuring validation across 3 industrial pilot sites and 3 dedicated testing zones with over six diverse use cases.
The project aims to deliver an open-source AI-powered system, suggesting a non-proprietary licensing model for the core software.
The system utilizes a cloud/edge computing split enabled by 5G connectivity to optimize trajectory and energy use.
The project period runs from 2022-10-01 to 2026-09-30, indicating it is currently in the development and piloting phase.
The consortium is heavily industry-driven, with 67% of the 9 partners coming from the industrial sector, including 3 SMEs. This high ratio of commercial partners across 4 countries (FR, ES, EL, CZ) suggests the project is focused on practical application rather than pure theory.
Contact CNRS in France for technical details on the open-source AI framework.
Contact us to find a partner for the 2026 deployment phase.
