Improving the Resilience of Quadrotors in Underground Environments by Combining Learning-based and Safety Controllers
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912949879177216 |
|---|---|
| author | Ward, Isaac Ronald Paral, Mark Riordan, Kristopher Kochenderfer, Mykel J. |
| author_facet | Ward, Isaac Ronald Paral, Mark Riordan, Kristopher Kochenderfer, Mykel J. |
| contents | Autonomously controlling quadrotors in large-scale subterranean environments is applicable to many areas such as environmental surveying, mining operations, and search and rescue. Learning-based controllers represent an appealing approach to autonomy, but are known to not generalize well to `out-of-distribution' environments not encountered during training. In this work, we train a normalizing flow-based prior over the environment, which provides a measure of how far out-of-distribution the quadrotor is at any given time. We use this measure as a runtime monitor, allowing us to switch between a learning-based controller and a safe controller when we are sufficiently out-of-distribution. Our methods are benchmarked on a point-to-point navigation task in a simulated 3D cave environment based on real-world point cloud data from the DARPA Subterranean Challenge Final Event Dataset. Our experimental results show that our combined controller simultaneously possesses the liveness of the learning-based controller (completing the task quickly) and the safety of the safety controller (avoiding collision). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_02808 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Improving the Resilience of Quadrotors in Underground Environments by Combining Learning-based and Safety Controllers Ward, Isaac Ronald Paral, Mark Riordan, Kristopher Kochenderfer, Mykel J. Robotics Artificial Intelligence Systems and Control Autonomously controlling quadrotors in large-scale subterranean environments is applicable to many areas such as environmental surveying, mining operations, and search and rescue. Learning-based controllers represent an appealing approach to autonomy, but are known to not generalize well to `out-of-distribution' environments not encountered during training. In this work, we train a normalizing flow-based prior over the environment, which provides a measure of how far out-of-distribution the quadrotor is at any given time. We use this measure as a runtime monitor, allowing us to switch between a learning-based controller and a safe controller when we are sufficiently out-of-distribution. Our methods are benchmarked on a point-to-point navigation task in a simulated 3D cave environment based on real-world point cloud data from the DARPA Subterranean Challenge Final Event Dataset. Our experimental results show that our combined controller simultaneously possesses the liveness of the learning-based controller (completing the task quickly) and the safety of the safety controller (avoiding collision). |
| title | Improving the Resilience of Quadrotors in Underground Environments by Combining Learning-based and Safety Controllers |
| topic | Robotics Artificial Intelligence Systems and Control |
| url | https://arxiv.org/abs/2509.02808 |