kth.sePublications KTH
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Towards Safe Reinforcement Learning with Reduced Conservativeness: A Case Study on Drone Flight Control
KTH, School of Electrical Engineering and Computer Science (EECS), Robotics, Perception and Learning.ORCID iD: 0000-0002-4633-6787
KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Robotics, Perception and Learning, RPL.ORCID iD: 0000-0003-3827-3824
University of Copenhagen, Department of Computer Science, Universitetsparken 1 Copenhagen, Denmark.
KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Robotics, Perception and Learning, RPL.ORCID iD: 0000-0003-2965-2953
2025 (English)In: IROS 2025 - 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems, Conference Proceedings, Institute of Electrical and Electronics Engineers (IEEE) , 2025, p. 14870-14876Conference paper, Published paper (Refereed)
Abstract [en]

Incorporating formal methods into reinforcement learning (RL) has the potential to result in the best of both worlds, combining the robustness of formal guarantees with the adaptability and learning capabilities of RL, though careful design is needed to balance safety and exploration. In this work, we propose a framework to mitigate this loss of exploration while still allowing for the safety of the system to be ensured. Specifically, we introduce a less restrictive method that can reduce the conservativeness of formal methods by refining a disturbance model using online collected data and it evaluates the safety of a learning-based controller, using computationally efficient zonotopic reachability analysis for the safety analysis to facilitate a real-time implementation. We validate the framework in a real-world drone flight through a canyon, where the drone is subjected to unknown external disturbances and the framework is tasked with learning those disturbances online and adjusting the safety guarantees accordingly. The results show that the framework enables a less restrictive online training of learning-based controllers without compromising the safety of the system.

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE) , 2025. p. 14870-14876
National Category
Control Engineering Robotics and automation Vehicle and Aerospace Engineering
Identifiers
URN: urn:nbn:se:kth:diva-377815DOI: 10.1109/IROS60139.2025.11247738Scopus ID: 2-s2.0-105029957051OAI: oai:DiVA.org:kth-377815DiVA, id: diva2:2044728
Conference
2025 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2025, Hangzhou, China, Oct 19 2025 - Oct 25 2025
Note

Part of ISBN 9798331543938

QC 20260310

Available from: 2026-03-10 Created: 2026-03-10 Last updated: 2026-03-10Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textScopus

Authority records

Hadjiloizou, LoizosWelle, Michael C.Kragic Jensfelt, Danica

Search in DiVA

By author/editor
Hadjiloizou, LoizosWelle, Michael C.Kragic Jensfelt, Danica
By organisation
Robotics, Perception and LearningRobotics, Perception and Learning, RPL
Control EngineeringRobotics and automationVehicle and Aerospace Engineering

Search outside of DiVA

GoogleGoogle Scholar

doi
urn-nbn

Altmetric score

doi
urn-nbn
Total: 87 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf