Data-Efficient Learning for Sim-to-Real Robotic Grasping using Deep Point Cloud Prediction Networks

Training a deep network policy for robot manipulation is notoriously costly and time consuming as it depends on collecting a significant amount of real world data. To work well in the real world, the policy needs to see many instances of the task, including various object arrangements in the scene as well as variations in object geometry, texture, material, and environmental illumination. In this paper, we propose a method that learns to perform table-top instance grasping of a wide variety of objects while using no real world grasping data, outperforming the baseline using 2.5D shape by 10%. Our method learns 3D point cloud of object, and use that to train a domain-invariant grasping policy. We formulate the learning process as a two-step procedure: 1) Learning a domain-invariant 3D shape representation of objects from about 76K episodes in simulation and about 530 episodes in the real world, where each episode lasts less than a minute and 2) Learning a critic grasping policy in simulation only based on the 3D shape representation from step 1. Our real world data collection in step 1 is both cheaper and faster compared to existing approaches as it only requires taking multiple snapshots of the scene using a RGBD camera. Finally, the learned 3D representation is not specific to grasping, and can potentially be used in other interaction tasks.

[1]  Rama Chellappa,et al.  Visual Domain Adaptation: A survey of recent advances , 2015, IEEE Signal Processing Magazine.

[2]  Jan Peters,et al.  Experiments with Hierarchical Reinforcement Learning of Multiple Grasping Policies , 2016, ISER.

[3]  Danica Kragic,et al.  Learning grasping points with shape context , 2010, Robotics Auton. Syst..

[4]  Gustavo Carneiro,et al.  Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue , 2016, ECCV.

[5]  Duc Thanh Nguyen,et al.  A Field Model for Repairing 3D Shapes , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  Kate Saenko,et al.  High precision grasp pose detection in dense clutter , 2016, 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[7]  Eren Erdal Aksoy,et al.  Part-based grasp planning for familiar objects , 2016, 2016 IEEE-RAS 16th International Conference on Humanoid Robots (Humanoids).

[8]  Danica Kragic,et al.  Dexterous grasping under shape uncertainty , 2016, Robotics Auton. Syst..

[9]  Silvio Savarese,et al.  DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[10]  Vittorio Ferrari,et al.  Learning to Generate and Reconstruct 3D Meshes with only 2D Supervision , 2018, BMVC.

[11]  Xiaojuan Qi,et al.  GAL: Geometric Adversarial Loss for Single-View 3D-Object Reconstruction , 2018, ECCV.

[12]  Ross B. Girshick,et al.  Mask R-CNN , 2017, 1703.06870.

[13]  Xinyu Liu,et al.  Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics , 2017, Robotics: Science and Systems.

[14]  Leonidas J. Guibas,et al.  PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Leonidas J. Guibas,et al.  ShapeNet: An Information-Rich 3D Model Repository , 2015, ArXiv.

[16]  Matthias Nießner,et al.  ScanComplete: Large-Scale Scene Completion and Semantic Segmentation for 3D Scans , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[17]  Lih-Yuan Deng,et al.  The Cross-Entropy Method: A Unified Approach to Combinatorial Optimization, Monte-Carlo Simulation, and Machine Learning , 2006, Technometrics.

[18]  Andrew W. Fitzgibbon,et al.  KinectFusion: Real-time dense surface mapping and tracking , 2011, 2011 10th IEEE International Symposium on Mixed and Augmented Reality.

[19]  J. Andrew Bagnell,et al.  Perceiving, learning, and exploiting object affordances for autonomous pile manipulation , 2013, Auton. Robots.

[20]  Subhransu Maji,et al.  Multiresolution Tree Networks for 3D Point Cloud Processing , 2018, ECCV.

[21]  Rob Fergus,et al.  Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-scale Convolutional Architecture , 2014, 2015 IEEE International Conference on Computer Vision (ICCV).

[22]  Thomas A. Funkhouser,et al.  Semantic Scene Completion from a Single Depth Image , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[23]  Honglak Lee,et al.  Perspective Transformer Nets: Learning Single-View 3D Object Reconstruction without 3D Supervision , 2016, NIPS.

[24]  Hao Li,et al.  Soft Rasterizer: Differentiable Rendering for Unsupervised Single-View Mesh Reconstruction , 2019, ArXiv.

[25]  Ales Leonardis,et al.  One-shot learning and generation of dexterous grasps for novel objects , 2016, Int. J. Robotics Res..

[26]  Edward H. Adelson,et al.  3D Shape Perception from Monocular Vision, Touch, and Shape Priors , 2018, 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[27]  Gabriela Csurka,et al.  Domain Adaptation for Visual Applications: A Comprehensive Survey , 2017, ArXiv.

[28]  Noah Snavely,et al.  Unsupervised Learning of Depth and Ego-Motion from Video , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  Peter K. Allen,et al.  Semantic grasping: planning task-specific stable robotic grasps , 2014, Auton. Robots.

[30]  Jiajun Wu,et al.  Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling , 2016, NIPS.

[31]  Honglak Lee,et al.  Deep learning for detecting robotic grasps , 2013, Int. J. Robotics Res..

[32]  Kate Saenko,et al.  Learning a visuomotor controller for real world robotic grasping using simulated depth images , 2017, CoRL.

[33]  P. Cochat,et al.  Et al , 2008, Archives de pediatrie : organe officiel de la Societe francaise de pediatrie.

[34]  Chad DeChant,et al.  Shape completion enabled robotic grasping , 2016, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[35]  Matei T. Ciocarlie,et al.  The Columbia grasp database , 2009, 2009 IEEE International Conference on Robotics and Automation.

[36]  Arkanath Pathak,et al.  Learning 6-DOF Grasping Interaction via Deep Geometry-Aware 3D Representations , 2018, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[37]  Abhinav Gupta,et al.  Supersizing self-supervision: Learning to grasp from 50K tries and 700 robot hours , 2015, 2016 IEEE International Conference on Robotics and Automation (ICRA).

[38]  Leonidas J. Guibas,et al.  Parsing Geometry Using Structure-Aware Shape Templates , 2018, 2018 International Conference on 3D Vision (3DV).

[39]  Sergey Levine,et al.  Grasp2Vec: Learning Object Representations from Self-Supervised Grasping , 2018, CoRL.

[40]  Hao Su,et al.  A Point Set Generation Network for 3D Object Reconstruction from a Single Image , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[41]  Wojciech Zaremba,et al.  Domain randomization for transferring deep neural networks from simulation to the real world , 2017, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[42]  Leonidas J. Guibas,et al.  Learning Representations and Generative Models for 3D Point Clouds , 2017, ICML.

[43]  Siddhartha S. Srinivasa,et al.  Physics-Based Grasp Planning Through Clutter , 2012, Robotics: Science and Systems.

[44]  Kate Saenko,et al.  Learning a visuomotor controller for real world robotic grasping using easily simulated depth images , 2017, ArXiv.

[45]  Sergey Levine,et al.  Sim-To-Real via Sim-To-Sim: Data-Efficient Robotic Grasping via Randomized-To-Canonical Adaptation Networks , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[46]  Sergey Levine,et al.  Learning hand-eye coordination for robotic grasping with deep learning and large-scale data collection , 2016, Int. J. Robotics Res..

[47]  Ozan ARSLAN.  3d Object Reconstruction from a Single Image. , 2014 .

[48]  Stefan Leutenegger,et al.  Deep learning a grasp function for grasping under gripper pose uncertainty , 2016, 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[49]  Tatsuya Harada,et al.  Neural 3D Mesh Renderer , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[50]  Sergey Levine,et al.  Using Simulation and Domain Adaptation to Improve Efficiency of Deep Robotic Grasping , 2018, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[51]  Wei Liu,et al.  Pixel2Mesh: Generating 3D Mesh Models from Single RGB Images , 2018, ECCV.

[52]  Koray Kavukcuoglu,et al.  Neural scene representation and rendering , 2018, Science.

[53]  Leonidas J. Guibas,et al.  Database‐Assisted Object Retrieval for Real‐Time 3D Reconstruction , 2015, Comput. Graph. Forum.

[54]  Stefan Ulbrich,et al.  OpenGRASP: A Toolkit for Robot Grasping Simulation , 2010, SIMPAR.

[55]  Mrinal Kalakrishnan,et al.  Multi-Task Domain Adaptation for Deep Learning of Instance Grasping from Simulation , 2017, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[56]  Ville Kyrki,et al.  Category-based task specific grasping , 2015, Robotics Auton. Syst..

[57]  T. Albright.  Perceiving , 2015, Daedalus.

[58]  Ashutosh Saxena,et al.  Robotic Grasping of Novel Objects using Vision , 2008, Int. J. Robotics Res..

[59]  Manuel Lopes,et al.  Active learning of visual descriptors for grasping using non-parametric smoothed beta distributions , 2012, Robotics Auton. Syst..

[60]  Danfei Xu,et al.  PointFusion: Deep Sensor Fusion for 3D Bounding Box Estimation , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[61]  Andrew J. Davison,et al.  Transferring End-to-End Visuomotor Control from Simulation to Real World for a Multi-Stage Task , 2017, CoRL.