Researchers from 10 countries, including Sveučilište Mondragon, collaborated on a project combining robotics, artificial intelligence, and software development. Among the 13 practical use cases, Mondragon Unibertsitatea focused on improving a manual refrigerator dismantling process performed by workers with disabilities, which posed risks of long-term muscle strain.
The team developed a simulation environment for a robot-human collaborative system . In this process, a conveyor belt aligns the refrigerator, the worker opens it, and the robot uses AI-powered vision to identify and remove the door’s magnetic seal while the worker removes internal components. This system ensures smooth coordination, with the conveyor moving the appliance to the next station after the task is completed.
A key innovation is the real-time monitoring system that prevents collisions between the robot and the worker. Using reinforcement learning, the control algorithms were tested in a digital validation environment to ensure safety before implementation. The project’s outcomes, including all use cases, are now available on its online platform, offering insights into safer and more efficient collaborative systems.
