Robotics Dexterous Grasping: The Methods Based on Point Cloud and Deep Learning.

Haonan Duan,Xiaofei Shen,Guangyun Xu,Wei Wei,Peng Wang,Yayu Huang

doi:10.3389/fnbot.2021.658280

Abstract

Dexterous manipulation, especially dexterous grasping, is a primitive and crucial ability of robots that allows the implementation of performing human-like behaviors. Deploying the ability on robots enables them to assist and substitute human to accomplish more complex tasks in daily life and industrial production. A comprehensive review of the methods based on point cloud and deep learning for robotics dexterous grasping from three perspectives is given in this paper. As a new category schemes of the mainstream methods, the proposed generation-evaluation framework is the core concept of the classification. The other two classifications based on learning modes and applications are also briefly described afterwards. This review aims to afford a guideline for robotics dexterous grasping researchers and developers.

Highlights

In the last decades, there has been an enormous proliferation in robotic community, both at in terms of research and attracting boundless varieties of imagination of general public, due to its diverse possibilities
Unlike the works mentioned above, this paper focuses on the methods of robotic dexterous grasping based on point cloud and deep learning
The current researches on robot dexterous grasp learning based on point cloud and deep learning can be divided into grasp candidate generation and grasp candidate evaluation

Summary

Introduction

There has been an enormous proliferation in robotic community, both at in terms of research and attracting boundless varieties of imagination of general public, due to its diverse possibilities. The vast majority of robots in operation today consist of 6 degree of freedom (6-DOF) which are either rotary (articulated) or sliding (prismatic), with a simple end effector for interacting with the workpieces (Murray et al, 1994). Robot manipulation means it can use and control different objects according to certain specifications and essentials through the end effector to achieve the effect of making the best use of playing the role of object itself (Okamura et al, 2000; Saut et al, 2007). Dexterous grasping is able to determine which posture to be employed to grasp where of the object to ensure a higher grasping success rate (Ciocarlie et al, 2007; Prattichizzo et al, 2008; Ciocarlie and Allen, 2009)

Objectives

Methods

Conclusion