國立虎尾科技大學 |

Antipodal Robotic Grasping using Deep Learning.

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	Antipodal Robotic Grasping using Deep Learning./
作者:	Joshi, Shirin.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2020,
面頁冊數:	72 p.
附註:	Source: Masters Abstracts International, Volume: 82-03.
Contained By:	Masters Abstracts International82-03.
標題:	Artificial intelligence. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28088105
ISBN:	9798664739336

Antipodal Robotic Grasping using Deep Learning.
Joshi, Shirin.

Antipodal Robotic Grasping using Deep Learning. - Ann Arbor : ProQuest Dissertations & Theses, 2020 - 72 p.

Source: Masters Abstracts International, Volume: 82-03.

Thesis (M.S.)--Rochester Institute of Technology, 2020.

This item must not be sold to any third party vendors.

In this work, we discuss two implementations that predict antipodal grasps for novel objects: A deep Q-learning approach and a Generative Residual Convolutional Neural Network approach. We present a deep reinforcement learning based method to solve the problem of robotic grasping using visio-motor feedback. The use of a deep learning based approach reduces the complexity caused by the use of hand-designed features. Our method uses an off-policy reinforcement learning framework to learn the grasping policy. We use the double deep Q-learning framework along with a novel Grasp-Q-Network to output grasp probabilities used to learn grasps that maximize the pick success. We propose a visual servoing mechanism that uses a multi-view camera setup that observes the scene which contains the objects of interest. We performed experiments using a Baxter Gazebo simulated environment as well as on the actual robot. The results show that our proposed method outperforms the baseline Q-learning framework and increases grasping accuracy by adapting a multi-view model in comparison to a single-view model. The second method tackles the problem of generating antipodal robotic grasps for unknown objects from an n-channel image of the scene. We propose a novel Generative Residual Convolutional Neural Network (GR-ConvNet) model that can generate robust antipodal grasps from n-channel input at real-time speeds (20ms). We evaluate the proposed model architecture on standard dataset and previously unseen household objects. We achieved state-of-the-art accuracy of 97.7% on Cornell grasp dataset. We also demonstrate a 93.5% grasp success rate on previously unseen real-world objects.

ISBN: 9798664739336Subjects--Topical Terms:

559380
Artificial intelligence.
Subjects--Index Terms:

Antipodal grasping

Antipodal Robotic Grasping using Deep Learning.
LDR:02829nam a2200373 4500 001 1038023
005 20210910100657.5
008 211029s2020 ||||||||||||||||| ||eng d
020 $a 9798664739336
035 $a (MiAaPQ)AAI28088105
035 $a AAI28088105
040 $a MiAaPQ $c MiAaPQ
100 1 $a Joshi, Shirin. $3 1335354
245 1 0 $a Antipodal Robotic Grasping using Deep Learning.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 72 p.
500 $a Source: Masters Abstracts International, Volume: 82-03.
500 $a Advisor: Sahin, Ferat.
502 $a Thesis (M.S.)--Rochester Institute of Technology, 2020.
506 $a This item must not be sold to any third party vendors.
520 $a In this work, we discuss two implementations that predict antipodal grasps for novel objects: A deep Q-learning approach and a Generative Residual Convolutional Neural Network approach. We present a deep reinforcement learning based method to solve the problem of robotic grasping using visio-motor feedback. The use of a deep learning based approach reduces the complexity caused by the use of hand-designed features. Our method uses an off-policy reinforcement learning framework to learn the grasping policy. We use the double deep Q-learning framework along with a novel Grasp-Q-Network to output grasp probabilities used to learn grasps that maximize the pick success. We propose a visual servoing mechanism that uses a multi-view camera setup that observes the scene which contains the objects of interest. We performed experiments using a Baxter Gazebo simulated environment as well as on the actual robot. The results show that our proposed method outperforms the baseline Q-learning framework and increases grasping accuracy by adapting a multi-view model in comparison to a single-view model. The second method tackles the problem of generating antipodal robotic grasps for unknown objects from an n-channel image of the scene. We propose a novel Generative Residual Convolutional Neural Network (GR-ConvNet) model that can generate robust antipodal grasps from n-channel input at real-time speeds (20ms). We evaluate the proposed model architecture on standard dataset and previously unseen household objects. We achieved state-of-the-art accuracy of 97.7% on Cornell grasp dataset. We also demonstrate a 93.5% grasp success rate on previously unseen real-world objects.
590 $a School code: 0465.
650 4 $a Artificial intelligence. $3 559380
650 4 $a Robotics. $3 561941
653 $a Antipodal grasping
653 $a Convolutional neural network
653 $a Deep learning
653 $a Deep reinforcement learning
653 $a Machine learning
653 $a Robotic grasping
690 $a 0771
690 $a 0800
710 2 $a Rochester Institute of Technology. $b Electrical Engineering. $3 1184409
773 0 $t Masters Abstracts International $g 82-03.
790 $a 0465
791 $a M.S.
792 $a 2020
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28088105