analiza aiqbench
SafeLoop introduces a non-invasive safety wrapper for Vision-Language-Action models that significantly reduces irreversible failures like collisions and object drops without retraining the base policy. By predicting hazard probabilities from vision and proprioception, the system dynamically decides to continue, checkpoint, or rollback to a safe state, improving the safety-success trade-off in both simulation and real-robot tasks. This approach addresses critical fragility issues in long-horizon robotic manipulation by adding proactive risk anticipation to existing VLA architectures.
SafeLoop: Risk-Aware Rollback for Vision-Language-Action Manipulation
arXiv:2609.26313v1 Announce Type: new Abstract: Recent vision-language-action (VLA) models are promising for general-purpose manipulation, but long-horizon execution remains fragile. Small state-estimation or control errors can lead to irreversible failures (e.g., collisions and object drops). Avoiding these risks requires a proactive safety mechanism capable of anticipating hazards. In this paper, we introduce SafeLoop, a non-invasive external wrapper that adds hazard prediction and rollback-based recovery to a VLA model without changing its parameters. SafeLoop trains a risk predictor from vision and proprioception to output four values: the probability and time-to-hazard for body collisions and for object failures. A lightweight controller then chooses one of three actions based on the predicted risk: continue execution (noop), save a safety checkpoint (record), or retreat in joint space (rollback). Rollback moves the robot back to a recent safe waypoint and queries the base policy again, which may yield an alternative continuation. Across 24 LIBERO tasks (16 random seeds each) and three real-robot tasks (25 rollouts each), SafeLoop achieves a stronger overall safety-success trade-off than alternative methods, reducing hazard cases by roughly 70% while preserving task success and the base-policy control rate. Project code is available at https://github.com/Loule0-0/SafeLoop/tree/release/safeloop.
Abstract:Recent vision-language-action (VLA) models are promising for general-purpose manipulation, but long-horizon execution remains fragile. Small state-estimation or control errors can lead to irreversible failures (e.g., collisions and object drops). Avoiding these risks requires a proactive safety mechanism capable of anticipating hazards. In this paper, we introduce SafeLoop, a non-invasive external wrapper that adds hazard prediction and rollback-based recovery to a VLA model without changing its parameters. SafeLoop trains a risk predictor from vision and proprioception to output four values: the probability and time-to-hazard for body collisions and for object failures. A lightweight controller then chooses one of three actions based on the predicted risk: continue execution (noop), save a safety checkpoint (record), or retreat in joint space (rollback). Rollback moves the robot back to a recent safe waypoint and queries the base policy again, which may yield an alternative continuation. Across 24 LIBERO tasks (16 random seeds each) and three real-robot tasks (25 rollouts each), SafeLoop achieves a stronger overall safety-success trade-off than alternative methods, reducing hazard cases by roughly 70% while preserving task success and the base-policy control rate. Project code is available at this https URL.
8 pages, 7 figures. Accepted to the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026)
Robotics (cs.RO)
arXiv:2609.26313 [cs.RO]
(or arXiv:2609.26313v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.26313 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 299 cuvinte
Subiecte conexe
Research Note: Filtering Subversion-Relevant Information From Pretraining Data Is Feasible
LessWrong · 5 oct. 2026
CR-VLA-Force: Learning Control-aware Compliance VLA Model for Robust Contact-rich Robotic Manipulation
arXiv · 9 sept. 2026
2026 BAIR Graduate Showcase
BAIR Blog · 1 iul. 2026
FabriVLA: A Lightweight Vision-Language-Action Model with Conformal Action Chunk Uncertainty
arXiv · 4 aug. 2026