Squeeze Grasping of Deformable Planar Objects with Segment Contacts and Stick/Slip Transitions

Feng Guo, Huan Lin, and Yan-Bin Jia

Abstract

Robotic grasping of a deformable object is difficult not simply due to the high computational cost of deformable modeling. More fundamentally, the difficulty lies in a wrench space that changes under deformation, with growing contact areas, and subject to varying slip/stick modes in these areas. This paper presents a grasping strategy by squeezing the object with two fingers. An analysis based on the finite element method (FEM) proves equilibrium and uniqueness of deformation during the action, and leads to a (improved) quadratic time deformation update from the displacements of as few as two contact nodes. An event-driven algorithm is then presented to track the contact regions during a squeeze, and determine the stick/slip mode of every node in contact. The contacts supply the constraints needed for deformation update using FEM. Several experiments with a Barrett Hand have been conducted for validation.