Salvato in:
Dettagli Bibliografici
Autori principali: Thakker, Rohan, Patnaik, Adarsh, Kurtz, Vince, Frey, Jonas, Becktor, Jonathan, Moon, Sangwoo, Royce, Rob, Kaufmann, Marcel, Georgakis, Georgios, Roth, Pascal, Burdick, Joel, Hutter, Marco, Khattak, Shehryar
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2506.17601
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911018575200256
author Thakker, Rohan
Patnaik, Adarsh
Kurtz, Vince
Frey, Jonas
Becktor, Jonathan
Moon, Sangwoo
Royce, Rob
Kaufmann, Marcel
Georgakis, Georgios
Roth, Pascal
Burdick, Joel
Hutter, Marco
Khattak, Shehryar
author_facet Thakker, Rohan
Patnaik, Adarsh
Kurtz, Vince
Frey, Jonas
Becktor, Jonathan
Moon, Sangwoo
Royce, Rob
Kaufmann, Marcel
Georgakis, Georgios
Roth, Pascal
Burdick, Joel
Hutter, Marco
Khattak, Shehryar
contents Safe, reliable navigation in extreme, unfamiliar terrain is required for future robotic space exploration missions. Recent generative-AI methods learn semantically aware navigation policies from large, cross-embodiment datasets, but offer limited safety guarantees. Inspired by human cognitive science, we propose a risk-guided diffusion framework that fuses a fast, learned "System-1" with a slow, physics-based "System-2", sharing computation at both training and inference to couple adaptability with formal safety. Hardware experiments conducted at the NASA JPL's Mars-analog facility, Mars Yard, show that our approach reduces failure rates by up to $4\times$ while matching the goal-reaching performance of learning-based robotic models by leveraging inference-time compute without any additional training.
format Preprint
id arxiv_https___arxiv_org_abs_2506_17601
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Risk-Guided Diffusion: Toward Deploying Robot Foundation Models in Space, Where Failure Is Not An Option
Thakker, Rohan
Patnaik, Adarsh
Kurtz, Vince
Frey, Jonas
Becktor, Jonathan
Moon, Sangwoo
Royce, Rob
Kaufmann, Marcel
Georgakis, Georgios
Roth, Pascal
Burdick, Joel
Hutter, Marco
Khattak, Shehryar
Robotics
Artificial Intelligence
Safe, reliable navigation in extreme, unfamiliar terrain is required for future robotic space exploration missions. Recent generative-AI methods learn semantically aware navigation policies from large, cross-embodiment datasets, but offer limited safety guarantees. Inspired by human cognitive science, we propose a risk-guided diffusion framework that fuses a fast, learned "System-1" with a slow, physics-based "System-2", sharing computation at both training and inference to couple adaptability with formal safety. Hardware experiments conducted at the NASA JPL's Mars-analog facility, Mars Yard, show that our approach reduces failure rates by up to $4\times$ while matching the goal-reaching performance of learning-based robotic models by leveraging inference-time compute without any additional training.
title Risk-Guided Diffusion: Toward Deploying Robot Foundation Models in Space, Where Failure Is Not An Option
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2506.17601