Robot safety research · standards · engineering
Robot Safety ForumROBOT SAFETY FORUM
한국어English简体中文
← Back to the library

Research paper

Linking reachability safety analysis with reinforcement learning

The paper reformulates safety analysis so reinforcement-learning methods can approximate safe sets and policies. The institutional abstract reports comparison with analytic/numerical solutions and simulated systems up to 18 states.

Reviewed 2026-11-05 · Source and bibliography checked

Bibliographic details and the editorial overview were checked against the original source. Access conditions and editions are identified with the source.
Original title
Bridging Hamilton-Jacobi Safety Analysis and Reinforcement Learning
Publisher
IEEE
Authors
Jaime F. Fisac, Neil F. Lugovoy, Vicenc Rubies-Royo, Shromona Ghosh, Claire J. Tomlin
DOI
10.1109/icra.2019.8794107
Conference date
2019-05
Source language
en
Edition
2019 International Conference on Robotics and Automation (ICRA)
Access
Publisher metadata and open author manuscript
Publication status
Published

Source overview

The paper reformulates safety analysis so reinforcement-learning methods can approximate safe sets and policies. The institutional abstract reports comparison with analytic/numerical solutions and simulated systems up to 18 states.

Use and review

Simulation evidence does not establish physical-robot safety. Model fidelity, constraint coverage and implementation effects require separate validation.

Sources & editions

Bridging Hamilton-Jacobi Safety Analysis and Reinforcement Learning ↗

IEEE · 2019 International Conference on Robotics and Automation (ICRA)

Primary author-institution abstract/DOI/authors and ICRA2019bibliography checked. Numerical/learned approximations do not automatically inherit exact reachability guarantees.