DISCOVER OUR INNOVATION PROJECTS

The PREHDIGIT project aims to demonstrate the contribution of multi-finger modular grippers that can adapt to different loads and shapes, to automate and allow the use of cobots in some industrial manipulation and handling tasks. By combining advanced mechatronics and teaching strategies, these systems aim to optimise productivity, heighten safety and improve quality, while making operators’ work less strenuous. The project focuses on the validation of their modularity and compliance with current manufacturing requirements.

The COGEFLUX project has led to the definition and validation of a method for optimising thermal comfort on the basis of energy consumption data from production workshops. The three use cases selected for control and prediction via an energy digital twin brought out valuable potential energy savings (of about 10%).

As part of the FactoryLab INTUIPROG project, the CEA laboratory (SRI/LCSR) has developed an intuitive and interactive programming solution based on the use of motion capture combined with assistance functions to make teaching easier, and allow repetitive tasks to be handled automatically by robots. This solution is compatible with different robot makes and models through a standardisation layer. The INTUIPROG solution allows operators and non-robotics specialists to define and start small robotic assistance tasks online. It takes the form of a software suite named SPIRE2, associated with motion capture hardware.

The PRECINET feasibility study is aimed at testing the approach of using artificial intelligence to detect pollution (grease and loose parts) in industrial piping, along with the possibility of incorporating AI into a video endoscope available in the market.

The project involved the creation of a tool that can measure the interaction between a human hand and the environment.
It led to the development of the multiphysics instrumentation of a glove and tests in operating conditions.

As part of the RACOSOUD project, high-performance tools for data processing, including those incorporating AI-based models, have been used in order to use the physical measurements collected during the welding process. The study was devoted to the resistance welding process.