Reinforcement Learning-based Control Allocation for the Innovative Control Effectors Aircraft

Pieter Simke de Vries; Erik-jan van Kampen

doi:10.2514/6.2019-0144

Reinforcement Learning-based Control Allocation for the Innovative Control Effectors Aircraft

Pieter Simke de Vries, Erik-jan van Kampen

Control & Simulation

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

8 Citations (Scopus)

280 Downloads (Pure)

Abstract

Established Control Allocation (CA) methods rely on knowledge of the control effectiveness for distributing control effector utilization for control of (overactuated) systems. The Innovative Control Effectors (ICE) aircraft model is highly overactuated with its 13 control effectors, CA is a preferred method to distribute control effector utilization. In this paper it is envisioned to use Reinforcement Learning (RL) for distributing control effector utilization, which requires no knowledge of the control effectiveness. RL allows to pursue more abstract and timescale separated objectives. The ICE aircraft’s altitude, considering only longitudinal motion, is controlled by distributing the control effector utilization using RL, from an initial offset, while pursuing secondary objectives such as decreasing effector utilization and thrust.

Original language	English
Title of host publication	AIAA Scitech 2019 Forum
Subtitle of host publication	7-11 January 2019, San Diego, California, USA
Number of pages	22
ISBN (Electronic)	978-1-62410-578-4
DOIs	https://doi.org/10.2514/6.2019-0144
Publication status	Published - 2019
Event	AIAA Scitech Forum, 2019 - San Diego, United States Duration: 7 Jan 2019 → 11 Jan 2019 https://arc.aiaa.org/doi/book/10.2514/MSCITECH19

Conference

Conference	AIAA Scitech Forum, 2019
Country/Territory	United States
City	San Diego
Period	7/01/19 → 11/01/19
Internet address	https://arc.aiaa.org/doi/book/10.2514/MSCITECH19

Access to Document

10.2514/6.2019-0144

6.2019-0144Final published version, 1.05 MB

Cite this

@inproceedings{124ef93316264c1f8cdfc592f7decac4,

title = "Reinforcement Learning-based Control Allocation for the Innovative Control Effectors Aircraft",

abstract = "Established Control Allocation (CA) methods rely on knowledge of the control effectiveness for distributing control effector utilization for control of (overactuated) systems. The Innovative Control Effectors (ICE) aircraft model is highly overactuated with its 13 control effectors, CA is a preferred method to distribute control effector utilization. In this paper it is envisioned to use Reinforcement Learning (RL) for distributing control effector utilization, which requires no knowledge of the control effectiveness. RL allows to pursue more abstract and timescale separated objectives. The ICE aircraft{\textquoteright}s altitude, considering only longitudinal motion, is controlled by distributing the control effector utilization using RL, from an initial offset, while pursuing secondary objectives such as decreasing effector utilization and thrust. ",

author = "{de Vries}, {Pieter Simke} and {van Kampen}, Erik-jan",

year = "2019",

doi = "10.2514/6.2019-0144",

language = "English",

booktitle = "AIAA Scitech 2019 Forum",

note = "AIAA Scitech Forum, 2019 ; Conference date: 07-01-2019 Through 11-01-2019",

url = "https://arc.aiaa.org/doi/book/10.2514/MSCITECH19",

}

TY - GEN

T1 - Reinforcement Learning-based Control Allocation for the Innovative Control Effectors Aircraft

AU - de Vries, Pieter Simke

AU - van Kampen, Erik-jan

PY - 2019

Y1 - 2019

N2 - Established Control Allocation (CA) methods rely on knowledge of the control effectiveness for distributing control effector utilization for control of (overactuated) systems. The Innovative Control Effectors (ICE) aircraft model is highly overactuated with its 13 control effectors, CA is a preferred method to distribute control effector utilization. In this paper it is envisioned to use Reinforcement Learning (RL) for distributing control effector utilization, which requires no knowledge of the control effectiveness. RL allows to pursue more abstract and timescale separated objectives. The ICE aircraft’s altitude, considering only longitudinal motion, is controlled by distributing the control effector utilization using RL, from an initial offset, while pursuing secondary objectives such as decreasing effector utilization and thrust.

AB - Established Control Allocation (CA) methods rely on knowledge of the control effectiveness for distributing control effector utilization for control of (overactuated) systems. The Innovative Control Effectors (ICE) aircraft model is highly overactuated with its 13 control effectors, CA is a preferred method to distribute control effector utilization. In this paper it is envisioned to use Reinforcement Learning (RL) for distributing control effector utilization, which requires no knowledge of the control effectiveness. RL allows to pursue more abstract and timescale separated objectives. The ICE aircraft’s altitude, considering only longitudinal motion, is controlled by distributing the control effector utilization using RL, from an initial offset, while pursuing secondary objectives such as decreasing effector utilization and thrust.

UR - http://www.scopus.com/inward/record.url?scp=85068942044&partnerID=8YFLogxK

U2 - 10.2514/6.2019-0144

DO - 10.2514/6.2019-0144

M3 - Conference contribution

BT - AIAA Scitech 2019 Forum

T2 - AIAA Scitech Forum, 2019

Y2 - 7 January 2019 through 11 January 2019

ER -

Reinforcement Learning-based Control Allocation for the Innovative Control Effectors Aircraft

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this