Visionbased computation of three-finger grasps on unknown planar objects

This paper presents an implemented vision-based strategy for computing three-finger stable grasps on unknown planar objects. Its only input is an im-age of a real unknown curved object instead of syn-thetic polygonal models. Grasp regions are identified for achieving tolerance around contact points so that errors in finger positioning as well as a finite fin-ger size are considered. It can find expanding and squeezing rasps involving internal and external re-gions, including holes. Theoretical constraints for force-closure are met while assuring robustness and real-time performance under real-world conditions. 1 In t roduct ion Extensive research in robotic grasping and dexter