The automation of production processes requires increasingly accurate and sensitive control of product handling. This ensures that the finished product is not damaged and that machines can adaptively handle process variations. In order to achieve the ability to manipulate an object as sensitively as a hand, it is necessary to integrate various technologies that enable spatial identification of the object and the gripping point, position robot and manipulator, and manipulate parts with a controlled level of force. The project aims to create a system for robotic manipulation of delicate products without
programming operations. To teach the gripping of a new part, it will be sufficient for the operator to perform part manipulation once, wearing a sensorized glove that can read the hand movement and the pressure exerted on the part. An AI algorithm will determine the ideal grip and ideal pressure to be applied by cross-referencing the object's shape and orientation data from the vision and from the glove worn by the operator. Once trained, the system will be able to replicate the operation automatically by adapting to process variability.