Modelling of the moulding process of a granular-based vacuum gripper with DEM

Abstract

Granular grippers are a promising approach to the flexible handling in soft robotics. As a result of the used granular materials, these grippers can grasp a wide spectrum of objects with many different shapes, especially compared to conventional mechanical or suction cup grippers. However, accurately predicting the graspability of differently shaped objects remains a challenge. Additionally, a comprehensive understanding of the various influences within the grasping mechanism is still lacking. Therefore, a specific granular-based gripper combining the principles of jamming and vacuum grippers was previously experimentally investigated for different object shapes, while varying various design parameters. In this study, the previous work is expanded through numerically modelling this specific gripper. For this purpose, the first sequence of the grasping process (moulding process) is modelled using the discrete element method, while the bonded particle method is used to model the membrane behaviour. The simulation shows good agreement with the experimental moulding results of differently shaped objects through optical comparisons. Furthermore, the parameters characterising the moulding are compared with a previously introduced object characteristic parameter, enabling the identification and characterisation of influences within the grasping mechanism.