Learning Pregrasp Manipulation of Objects from Ungraspable Poses

2020/02/15 by Zhaole Sun, Kai Yuan, Sun, Zhaole +7 · 3 citations
Computer Science · Engineering · Neuroscience · #Artificial intelligence #Computer science #Computer vision #FOS: Computer and information sciences #GRASP #Human–computer interaction #Lift (data mining) #Machine learning #Motor Control and Adaptation #Muscle activation and electromyography studies #Object (grammar) #Reinforcement learning #Robot #Robot Manipulation and Learning #Robotics (cs.RO) #Table (database) #cs.RO

paper · pdf · doi:10.48550/arxiv.2002.06344

published in arXiv (Cornell University) (Cornell University) · 8 pages open access version for ICRA2020 6 pages acceptance paper

arxiv created 2020/02/15 · openalex publication_date 2020/02/15 · arxiv updated 2020/02/18 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08

Abstract

In robotic grasping, objects are often occluded in ungraspable configurations such that no pregrasp pose can be found, eg large flat boxes on the table that can only be grasped from the side. Inspired by humans' bimanual manipulation, eg one hand to lift up things and the other to grasp, we address this type of problems by introducing pregrasp manipulation - push and lift actions. We propose a model-free Deep Reinforcement Learning framework to train control policies that utilize visual information and proprioceptive states of the robot to autonomously discover robust pregrasp manipulation. The robot arm learns to first push the object towards a support surface and establishes a pivot to lift up one side of the object, thus creating a clearance between the object and the table for possible grasping solutions. Furthermore, we show the effectiveness of our proposed learning framework in training robust pregrasp policies that can directly transfer from simulation to real hardware through suitable design of training procedures, state, and action space. Lastly, we evaluate the effectiveness and the generalisation ability of the learned policies in real-world experiments, and demonstrate pregrasp manipulation of objects with various size, shape, weight, and surface friction.

Citations

• Adam: A Method for Stochastic Optimization
• Playing Atari with Deep Reinforcement Learning
• Continuous control with deep reinforcement learning
• Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
• Learning Dexterous In-Hand Manipulation
• Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics

Cited by

• End-to-End Dexterous Arm-Hand VLA Policies via Shared Autonomy: VR Teleoperation Augmented by Autonomous Hand VLA Policy for Efficient Data Collection
• Leveraging Extrinsic Dexterity for Occluded Grasping on Grasp Constraining Walls
• Learning Dual-Arm Coordination for Grasping Large Flat Objects

Related

• Reinforcement Learning Based Pushing and Grasping Objects from Ungraspable Poses
• Learning Synergies between Pushing and Grasping with Self-supervised Deep Reinforcement Learning
• Learning Extrinsic Dexterity with Parameterized Manipulation Primitives
• Towards Real-World Efficiency: Domain Randomization in Reinforcement Learning for Pre-Capture of Free-Floating Moving Targets by Autonomous Robots
• Learning Dual-Arm Coordination for Grasping Large Flat Objects