Imagine a robot that learns to pick up things just by watching a human do it, rather than being programmed for every single move. It uses a sense of touch, like human fingertips, to feel if an object is slipping or too soft. This allows the machine to handle tricky items like wires or fresh food without breaking them. It's essentially giving robots a 'brain' that learns from its mistakes in real-time.
Robots struggle to handle soft, flexible, or unknown objects because they lack the tactile sense and adaptability of humans. This prevents automation in sectors like food handling and complex electronics assembly.
An AI-powered manipulation system featuring flexible tactile sensors, a CNN-based object pose estimation system, and a ROS2-integrated control module for service robots.
If you are a packaging facility dealing with the fragility of fresh produce — this project developed a manipulation system that uses tactile sensors and AI to handle food without damage. This reduces waste and allows for automation of tasks previously done by hand.
If you are a prosthetics manufacturer dealing with the lack of sensory feedback in artificial limbs — this project developed flexible tactile sensors for the Hannes and AR10 hands. This allows users to measure grasping force more accurately, improving the utility of the prosthetic.
If you are an assembly plant dealing with the difficulty of automating wire insertion — this project developed a robotic cell using CNN-based position estimation. This increases the success rate of grasping and inserting flexible wires into components.
Based on available project data, the specific commercial pricing for the system is not provided, as the project was funded by a EUR 4,509,303 EU contribution for research and development.
The project tested the system in specific robotic cells for wire grasping and kitchen tasks, suggesting a path toward industrial scale, though full-scale deployment data is not detailed.
Based on available project data, the IP details and licensing terms are not specified in the summary; these would typically be managed by the 14 consortium partners.
The system is designed for flexibility, having been integrated with TIAGo service robots using ROS2 modules and various prosthetic hands.
The project period runs from 2022-09-01 to 2026-04-30, indicating that final validation and deliverables are expected by April 2026.
The consortium is heavily weighted toward academic research with 7 universities and 3 research institutes, reflecting the project's focus on AI development. However, the inclusion of 2 industry partners and 1 SME (14% industry ratio) ensures that the developed tactile sensors and manipulation algorithms are tested against real-world hardware like the TIAGo robot and AR10 prosthetic hands.
Contact the AI and Robotics department at Alma Mater Studiorum - Universita di Bologna
Contact us to connect with the IntelliMan consortium for licensing tactile sensor technology.
