Safe Learning in the Real World via Adaptive Shielding with Hamilton-Jacobi Reachability

Abstract

We present a robust shielding framework using Hamilton-Jacobi Reachability that can be combined with any off-policy Reinforcement Learning to enable safer learning. Using an approximate model of a system dynamics, our method can capture the local model mismatch from a safety perspective. This leads to a more conservative safety filter that can adapt to model mismatch. Using a Turtlebot 2, we demonstrate that our method can allow for safe learning in the real-world with minimal human intervention.

Cite

Text

Lu et al. "Safe Learning in the Real World via Adaptive Shielding with Hamilton-Jacobi Reachability." Proceedings of the 7th Annual Learning for Dynamics \& Control Conference, 2025.

Markdown

[Lu et al. "Safe Learning in the Real World via Adaptive Shielding with Hamilton-Jacobi Reachability." Proceedings of the 7th Annual Learning for Dynamics \& Control Conference, 2025.](https://mlanthology.org/l4dc/2025/lu2025l4dc-safe/)

BibTeX

@inproceedings{lu2025l4dc-safe,
  title     = {{Safe Learning in the Real World via Adaptive Shielding with Hamilton-Jacobi Reachability}},
  author    = {Lu, Michael and Gosain, Jashanraj and Sang, Luna and Chen, Mo},
  booktitle = {Proceedings of the 7th Annual Learning for Dynamics \& Control Conference},
  year      = {2025},
  pages     = {1257-1270},
  volume    = {283},
  url       = {https://mlanthology.org/l4dc/2025/lu2025l4dc-safe/}
}