Research paper
Reachability supervision · Intervening as learning confidence changes
Hamilton–Jacobi analysis and Bayesian updates supervise a learning controller. The supervisor intervenes at safety boundaries or when new observations reduce confidence; the authors report quadrotor learning experiments with an introduced external disturbance.
Reviewed 2026-11-05 · Source and bibliography checked
- Original title
- A General Safety Framework for Learning-Based Control in Uncertain Robotic Systems
- Publisher
- Institute of Electrical and Electronics Engineers (IEEE)
- Authors
- Jaime F. Fisac, Anayo K. Akametalu, Melanie N. Zeilinger, Shahab Kaynama, Jeremy Gillula, Claire J. Tomlin
- DOI
- 10.1109/tac.2018.2876389
- Publication date
- 2019-07
- Source language
- en
- Edition
- IEEE Transactions on Automatic Control; 2019; DOI 10.1109/tac.2018.2876389
- Access
- Publisher metadata and open author manuscript
- Publication status
- Published
Related resources and engineering notes
Source overview
Hamilton–Jacobi analysis and Bayesian updates supervise a learning controller. The supervisor intervenes at safety boundaries or when new observations reduce confidence; the authors report quadrotor learning experiments with an introduced external disturbance.
Use and review
Checked author abstract and publication metadata. Guarantees depend on approximate dynamics and statistical/worst-case assumptions; limited flight experiments do not establish safety across all operating conditions.
Sources & editions
A General Safety Framework for Learning-Based Control in Uncertain Robotic Systems ↗
Institute of Electrical and Electronics Engineers (IEEE) · IEEE Transactions on Automatic Control; 2019; DOI 10.1109/tac.2018.2876389
IEEE TAC bibliography; matching accepted author manuscript abstract https://arxiv.org/abs/1705.01292v3. Journal date 2019-07 retained.