Abstract
We consider a novel problem where multiple rigid convex polygonal objectsrest in randomly placed positions and orientations on a planar surface visiblefrom an overhead camera. The objective is to efficiently grasp and transportall objects into a bin using multi-object push-grasps, where multiple objectsare pushed together to facilitate multi-object grasping. We provide necessaryconditions for frictionless multi-object push-grasps and apply these to filterinadmissible grasps in a novel multi-object grasp planner. We find that ourplanner is 19 times faster than a Mujoco simulator baseline. We also propose apicking algorithm that uses both single- and multi-object grasps to pickobjects. In physical grasping experiments comparing performance with asingle-object picking baseline, we find that the frictionless multi-objectgrasping system achieves 13.6\% higher grasp success and is 59.9\% faster, from212 PPH to 340 PPH. See\url{https://sites.google.com/view/multi-object-grasping} for videos and code.