An Efficient Deep Reinforcement Learning Framework for UAVs

Shanglin Zhou; Bingbing Li; Caiwu Ding; Lu Lu; Caiwen Ding

doi:10.1109/ISQED48828.2020.9136980

An Efficient Deep Reinforcement Learning Framework for UAVs

Shanglin Zhou, Bingbing Li, Caiwu Ding, Lu Lu, Caiwen Ding

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

14 Scopus citations

Abstract

3D Dynamic simulator such as Gazebo has become a popular substitution for unmanned aerial vehicle (UAV) because of its user-friendly in real-world scenarios. At this point, well-functioning algorithms on the UAV controller are needed for guidance, navigation, and control for autonomous navigation. Deep reinforcement learning (DRL) comes into sight as its famous self-learning characteristic. This goal-orientated algorithm can learn how to attain a complex objective or maximize along a particular dimension over many steps. In this paper, we propose a general framework to incorporate DRL with the UAV simulation environment. The whole system consists of the DRL algorithm for attitude control, packing algorithm on the Robot Operation System (ROS) to connect DRL with PX4 controller, and a Gazebo simulator that emulates the real-world environment. Experimental results demonstrate the effectiveness of the proposed framework.

Original language	English (US)
Title of host publication	Proceedings of the 21st International Symposium on Quality Electronic Design, ISQED 2020
Publisher	IEEE Computer Society
Pages	323-328
Number of pages	6
ISBN (Electronic)	9781728142074
DOIs	https://doi.org/10.1109/ISQED48828.2020.9136980
State	Published - Mar 2020
Externally published	Yes
Event	21st International Symposium on Quality Electronic Design, ISQED 2020 - Santa Clara, United States Duration: Mar 25 2020 → Mar 26 2020

Publication series

Name	Proceedings - International Symposium on Quality Electronic Design, ISQED
Volume	2020-March
ISSN (Print)	1948-3287
ISSN (Electronic)	1948-3295

Conference

Conference	21st International Symposium on Quality Electronic Design, ISQED 2020
Country/Territory	United States
City	Santa Clara
Period	3/25/20 → 3/26/20

Bibliographical note

Publisher Copyright:
© 2020 IEEE.

Keywords

Deep Reinforcement Learning
Gazebo
PX4
Simulation Environment
Unmanned Aerial Vehicle

Publisher link

10.1109/ISQED48828.2020.9136980

Find It

Find It at U of M Libraries

Cite this

Zhou, S., Li, B., Ding, C., Lu, L., & Ding, C. (2020). An Efficient Deep Reinforcement Learning Framework for UAVs. In Proceedings of the 21st International Symposium on Quality Electronic Design, ISQED 2020 (pp. 323-328). Article 9136980 (Proceedings - International Symposium on Quality Electronic Design, ISQED; Vol. 2020-March). IEEE Computer Society. https://doi.org/10.1109/ISQED48828.2020.9136980

An Efficient Deep Reinforcement Learning Framework for UAVs. / Zhou, Shanglin; Li, Bingbing; Ding, Caiwu et al.
Proceedings of the 21st International Symposium on Quality Electronic Design, ISQED 2020. IEEE Computer Society, 2020. p. 323-328 9136980 (Proceedings - International Symposium on Quality Electronic Design, ISQED; Vol. 2020-March).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Zhou, S, Li, B, Ding, C, Lu, L & Ding, C 2020, An Efficient Deep Reinforcement Learning Framework for UAVs. in Proceedings of the 21st International Symposium on Quality Electronic Design, ISQED 2020., 9136980, Proceedings - International Symposium on Quality Electronic Design, ISQED, vol. 2020-March, IEEE Computer Society, pp. 323-328, 21st International Symposium on Quality Electronic Design, ISQED 2020, Santa Clara, United States, 3/25/20. https://doi.org/10.1109/ISQED48828.2020.9136980

@inproceedings{e9adc8cde3d74a98843466c5952fcdfc,

title = "An Efficient Deep Reinforcement Learning Framework for UAVs",

abstract = "3D Dynamic simulator such as Gazebo has become a popular substitution for unmanned aerial vehicle (UAV) because of its user-friendly in real-world scenarios. At this point, well-functioning algorithms on the UAV controller are needed for guidance, navigation, and control for autonomous navigation. Deep reinforcement learning (DRL) comes into sight as its famous self-learning characteristic. This goal-orientated algorithm can learn how to attain a complex objective or maximize along a particular dimension over many steps. In this paper, we propose a general framework to incorporate DRL with the UAV simulation environment. The whole system consists of the DRL algorithm for attitude control, packing algorithm on the Robot Operation System (ROS) to connect DRL with PX4 controller, and a Gazebo simulator that emulates the real-world environment. Experimental results demonstrate the effectiveness of the proposed framework.",

keywords = "Deep Reinforcement Learning, Gazebo, PX4, Simulation Environment, Unmanned Aerial Vehicle",

author = "Shanglin Zhou and Bingbing Li and Caiwu Ding and Lu Lu and Caiwen Ding",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 21st International Symposium on Quality Electronic Design, ISQED 2020 ; Conference date: 25-03-2020 Through 26-03-2020",

year = "2020",

month = mar,

doi = "10.1109/ISQED48828.2020.9136980",

language = "English (US)",

series = "Proceedings - International Symposium on Quality Electronic Design, ISQED",

publisher = "IEEE Computer Society",

pages = "323--328",

booktitle = "Proceedings of the 21st International Symposium on Quality Electronic Design, ISQED 2020",

}

TY - GEN

T1 - An Efficient Deep Reinforcement Learning Framework for UAVs

AU - Zhou, Shanglin

AU - Li, Bingbing

AU - Ding, Caiwu

AU - Lu, Lu

AU - Ding, Caiwen

PY - 2020/3

Y1 - 2020/3

N2 - 3D Dynamic simulator such as Gazebo has become a popular substitution for unmanned aerial vehicle (UAV) because of its user-friendly in real-world scenarios. At this point, well-functioning algorithms on the UAV controller are needed for guidance, navigation, and control for autonomous navigation. Deep reinforcement learning (DRL) comes into sight as its famous self-learning characteristic. This goal-orientated algorithm can learn how to attain a complex objective or maximize along a particular dimension over many steps. In this paper, we propose a general framework to incorporate DRL with the UAV simulation environment. The whole system consists of the DRL algorithm for attitude control, packing algorithm on the Robot Operation System (ROS) to connect DRL with PX4 controller, and a Gazebo simulator that emulates the real-world environment. Experimental results demonstrate the effectiveness of the proposed framework.

AB - 3D Dynamic simulator such as Gazebo has become a popular substitution for unmanned aerial vehicle (UAV) because of its user-friendly in real-world scenarios. At this point, well-functioning algorithms on the UAV controller are needed for guidance, navigation, and control for autonomous navigation. Deep reinforcement learning (DRL) comes into sight as its famous self-learning characteristic. This goal-orientated algorithm can learn how to attain a complex objective or maximize along a particular dimension over many steps. In this paper, we propose a general framework to incorporate DRL with the UAV simulation environment. The whole system consists of the DRL algorithm for attitude control, packing algorithm on the Robot Operation System (ROS) to connect DRL with PX4 controller, and a Gazebo simulator that emulates the real-world environment. Experimental results demonstrate the effectiveness of the proposed framework.

KW - Deep Reinforcement Learning

KW - Gazebo

KW - PX4

KW - Simulation Environment

KW - Unmanned Aerial Vehicle

UR - http://www.scopus.com/inward/record.url?scp=85089938371&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85089938371&partnerID=8YFLogxK

U2 - 10.1109/ISQED48828.2020.9136980

DO - 10.1109/ISQED48828.2020.9136980

M3 - Conference contribution

AN - SCOPUS:85089938371

T3 - Proceedings - International Symposium on Quality Electronic Design, ISQED

SP - 323

EP - 328

BT - Proceedings of the 21st International Symposium on Quality Electronic Design, ISQED 2020

PB - IEEE Computer Society

T2 - 21st International Symposium on Quality Electronic Design, ISQED 2020

Y2 - 25 March 2020 through 26 March 2020

ER -

An Efficient Deep Reinforcement Learning Framework for UAVs

Abstract

Publication series

Conference

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this