A self-adaptive high precision gripper for shape variant components: Towards higher reliability and efficiency of a cobotic cell

With continuously shrinking batch sizes and increasing order numbers the handling of different object variants becomes a key parameter for the automatization of assembly processes in Industry 4.0. Many compliant and underactuated grasping devices that exploit hierarchical principles, including whipple trees, tendon-pulley systems, pin arrays, fluid differentials, and air pressure, just to name a few, are subject to research. However, these often bionically inspired systems are mostly used for humanoid applications, such as prosthesis, dexterous hands, and crop harvesting. While they can achieve a significant grasp robustness, their precision is often too low for industrial manufacturing processes. Therefore, this paper proposes a deterministic grasping system with one level of adaptation to grasp a great number of object variants in a cobotic assembly cell. This grasping system not only adapts to different shapes but also compensates initial position and alignment errors and guarantees a predefined centring within the grasp. The technology has been tested intensively in the laboratory and subsequently demonstrated in the cobotic cell, where it decreased the machine breakdown time due to pick and place failures by 90%.