Towards Precise Robotic Grasping by Probabilistic Post-grasp Displacement Estimation

EasyChair Preprint 1282, version 2

Abstract

Precise robotic grasping is important for many industrial applications, such as assembly and palletizing, where the location of the object needs to be controlled and known. However, achieving precise grasps is challenging due to noise in sensing and control, as well as unknown object properties. We propose a method to plan robotic grasps that are both robust and precise by training two convolutional neural networks - one to predict robustness of a grasp and another to predict a distribution of post-grasp object displacements. Our networks are trained with depth images in simulation on a dataset of over 1000 industrial parts and were successfully deployed on a real robot without having to be further fine-tuned. The proposed displacement estimator achieves a mean prediction errors of 0.68cm and 3.42deg on novel objects in real world experiments. It also reduces the standard deviation of the translation prediction errors by a factor of x4.36 over baselines that do not optimize for grasp displacement variance. Supplementary material is available at: https://precise-grasping.jialiangz.me.

Keyphrases: Grasp Quality, Object displacement, Robotic Grasping, grasp displacement, grasp displacement estimation, grasp displacement prediction, grasp quality network, model learning, oliver kroemer, precise grasp, precise robotic grasping, probabilistic post grasp displacement, self-supervised learning

@booklet{EasyChair:1282,
  author    = {Jialiang Zhao and Jacky Liang and Oliver Kroemer},
  title     = {Towards Precise Robotic Grasping by Probabilistic Post-grasp Displacement Estimation},
  howpublished = {EasyChair Preprint 1282},
  year      = {EasyChair, 2019}}