Twin-Delayed Deep Deterministic Policy Gradient for altitude control of a flying-wing aircraft with an uncertain aerodynamic model

W.J.E. Völker, Y. Li, E. van Kampen

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

140 Downloads (Pure)

Abstract

Recent research on the Flying V - a flying-wing long-range passenger aircraft - shows that its airframe design is 25% more aerodynamically efficient than a conventional tube-and-wing airframe. The Flying V is therefore a promising contribution towards reduction in climate impact of long-haul flights. However, some design aspects of the Flying V still remain to be investigated, one of which is automatic flight control. Due to the unconventional airframe shape of the Flying V, aerodynamic modelling cannot rely on validated aerodynamic-modelling tools and the accuracy of the aerodynamic model is uncertain. Therefore, this contribution investigates how an automatic flight controller that is robust to aerodynamic-model uncertainty can be developed, by utilising Twin-Delayed Deep Deterministic Policy Gradient (TD3) - a recent deep-reinforcement-learning algorithm. The results show that an offline-trained single-loop altitude controller that is fully based on TD3 can track a given altitude-reference signal and is robust to aerodynamic-model uncertainty of more than 25%.
Original languageEnglish
Title of host publicationAIAA SciTech Forum 2023
Number of pages19
ISBN (Electronic)978-1-62410-699-6
DOIs
Publication statusPublished - 2023
EventAIAA SCITECH 2023 Forum - National Harbor, MD & Online, Washington, United States
Duration: 23 Jan 202327 Jan 2023
https://arc-aiaa-org.tudelft.idm.oclc.org/doi/book/10.2514/MSCITECH23

Conference

ConferenceAIAA SCITECH 2023 Forum
Country/TerritoryUnited States
CityWashington
Period23/01/2327/01/23
Internet address

Fingerprint

Dive into the research topics of 'Twin-Delayed Deep Deterministic Policy Gradient for altitude control of a flying-wing aircraft with an uncertain aerodynamic model'. Together they form a unique fingerprint.

Cite this