2022/01/31 by Anuj Pasricha, Yi-Shiuan Tung, Pasricha, Anuj +5 · 2 citations
Computer Science · Engineering · Medicine · Psychology · #Adaptability #Artificial intelligence #Computer science #Computer vision #Engineering #FOS: Computer and information sciences #Human–computer interaction #Medicine #Object (grammar) #Prehensile tail #Psychology #Reinforcement Learning in Robotics #Robot #Robot Manipulation and Learning #Robotic Path Planning Algorithms #Robotics (cs.RO) #Simulation #Task (project management) #Workspace
paper · pdf · doi:10.48550/arxiv.2201.13428
published in arXiv (Cornell University) (Cornell University)
openalex publication_date 2022/01/31 · openalex created_date 2022/07/24 · openalex updated_date 2026/08/06
In this work, we introduce PokeRRT, a novel motion planning algorithm that\ndemonstrates poking as an effective non-prehensile manipulation skill to enable\nfast manipulation of objects and increase the size of a robot's reachable\nworkspace. We showcase poking as a failure recovery tactic used synergistically\nwith pick-and-place for resiliency in cases where pick-and-place initially\nfails or is unachievable. Our experiments demonstrate the efficiency of the\nproposed framework in planning object trajectories using poking manipulation in\nuncluttered and cluttered environments. In addition to quantitatively and\nqualitatively demonstrating the adaptability of PokeRRT to different scenarios\nin both simulation and real-world settings, our results show the advantages of\npoking over pushing and grasping in terms of success rate and task time.\n