Perceive With Confidence: Statistical Safety Assurances for Navigation with Learning-Based Perception

Anushri Dixit, Zhiting Mei, Meghan Booker, Mariko Storey-Matsutani, Allen Z. Ren, Anirudha Majumdar
Proceedings of The 8th Conference on Robot Learning, PMLR 270:2517-2541, 2025.

Abstract

Rapid advances in perception have enabled large pre-trained models to be used out of the box for transforming high-dimensional, noisy, and partial observations of the world into rich occupancy representations. However, the reliability of these models and consequently their safe integration onto robots remains unknown when deployed in environments unseen during training. In this work, we address this challenge by rigorously quantifying the uncertainty of pre-trained perception systems for object detection via a novel calibration technique based on conformal prediction. Crucially, this procedure guarantees robustness to distribution shifts in states when perceptual outputs are used in conjunction with a planner. As a result, the calibrated perception system can be used in combination with any safe planner to provide an end-to-end statistical assurance on safety in unseen environments. We evaluate the resulting approach, Perceive with Confidence (PwC), with experiments in simulation and on hardware where a quadruped robot navigates through previously unseen indoor, static environments. These experiments validate the safety assurances for obstacle avoidance provided by PwC and demonstrate up to 40% improvements in empirical safety compared to baselines.

Cite this Paper

BibTeX
@InProceedings{pmlr-v270-dixit25a, title = {Perceive With Confidence: Statistical Safety Assurances for Navigation with Learning-Based Perception}, author = {Dixit, Anushri and Mei, Zhiting and Booker, Meghan and Storey-Matsutani, Mariko and Ren, Allen Z. and Majumdar, Anirudha}, booktitle = {Proceedings of The 8th Conference on Robot Learning}, pages = {2517--2541}, year = {2025}, editor = {Agrawal, Pulkit and Kroemer, Oliver and Burgard, Wolfram}, volume = {270}, series = {Proceedings of Machine Learning Research}, month = {06--09 Nov}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v270/main/assets/dixit25a/dixit25a.pdf}, url = {https://proceedings.mlr.press/v270/dixit25a.html}, abstract = {Rapid advances in perception have enabled large pre-trained models to be used out of the box for transforming high-dimensional, noisy, and partial observations of the world into rich occupancy representations. However, the reliability of these models and consequently their safe integration onto robots remains unknown when deployed in environments unseen during training. In this work, we address this challenge by rigorously quantifying the uncertainty of pre-trained perception systems for object detection via a novel calibration technique based on conformal prediction. Crucially, this procedure guarantees robustness to distribution shifts in states when perceptual outputs are used in conjunction with a planner. As a result, the calibrated perception system can be used in combination with any safe planner to provide an end-to-end statistical assurance on safety in unseen environments. We evaluate the resulting approach, Perceive with Confidence (PwC), with experiments in simulation and on hardware where a quadruped robot navigates through previously unseen indoor, static environments. These experiments validate the safety assurances for obstacle avoidance provided by PwC and demonstrate up to 40% improvements in empirical safety compared to baselines.} }
Endnote
%0 Conference Paper %T Perceive With Confidence: Statistical Safety Assurances for Navigation with Learning-Based Perception %A Anushri Dixit %A Zhiting Mei %A Meghan Booker %A Mariko Storey-Matsutani %A Allen Z. Ren %A Anirudha Majumdar %B Proceedings of The 8th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2025 %E Pulkit Agrawal %E Oliver Kroemer %E Wolfram Burgard %F pmlr-v270-dixit25a %I PMLR %P 2517--2541 %U https://proceedings.mlr.press/v270/dixit25a.html %V 270 %X Rapid advances in perception have enabled large pre-trained models to be used out of the box for transforming high-dimensional, noisy, and partial observations of the world into rich occupancy representations. However, the reliability of these models and consequently their safe integration onto robots remains unknown when deployed in environments unseen during training. In this work, we address this challenge by rigorously quantifying the uncertainty of pre-trained perception systems for object detection via a novel calibration technique based on conformal prediction. Crucially, this procedure guarantees robustness to distribution shifts in states when perceptual outputs are used in conjunction with a planner. As a result, the calibrated perception system can be used in combination with any safe planner to provide an end-to-end statistical assurance on safety in unseen environments. We evaluate the resulting approach, Perceive with Confidence (PwC), with experiments in simulation and on hardware where a quadruped robot navigates through previously unseen indoor, static environments. These experiments validate the safety assurances for obstacle avoidance provided by PwC and demonstrate up to 40% improvements in empirical safety compared to baselines.
APA
Dixit, A., Mei, Z., Booker, M., Storey-Matsutani, M., Ren, A.Z. & Majumdar, A.. (2025). Perceive With Confidence: Statistical Safety Assurances for Navigation with Learning-Based Perception. Proceedings of The 8th Conference on Robot Learning, in Proceedings of Machine Learning Research 270:2517-2541 Available from https://proceedings.mlr.press/v270/dixit25a.html.

Related Material