aiqbench analysis
WBAG addresses a critical gap in Vision-Language-Action (VLA) robotics by integrating whole-body and attached-object geometry into safety constraints, moving beyond simplified end-effector models to prevent collisions with the full articulated system. By converting these evolving geometric representations into differentiable Control Barrier Function constraints, the framework minimally modifies native VLA actions while achieving superior scene-level safety metrics on the SafeLIBERO benchmark. This advancement is essential for deploying generalizable robotic manipulation policies in unstructured real-world environments where complex kinematic interactions are common.
WBAG: A Whole-Body and Attached-Geometry Safety Framework for Vision-Language-Action Manipulation
arXiv:2610.01083v1 Announce Type: new Abstract: Vision-language-action (VLA) policies have demonstrated impressive capabilities in generalizable robotic manipulation, but their deployment in the real world remains challenging due to potential collisions involving different parts of the robot, manipulated objects, and the surrounding environment. Existing inference-time VLA safety frameworks typically rely on simplified end-effector-centered representations that do not explicitly model the full articulated robot and attached-object geometry. In this paper, we present WBAG, a safety framework that models the robot's whole-body and grasp-dependent attached geometry. WBAG constructs a grasp-conditioned safe set that adapts the protected geometry as objects are grasped, then converts this evolving geometry into differentiable CBF constraints that minimally modify the VLA's native six-dimensional operational-space action for collision avoidance across robot, scene, and attached geometry. On the SafeLIBERO benchmark, a variant of LIBERO augmented with obstacles for safety evaluation, WBAG achieves the best overall safety and safe task success among the evaluated methods under a scene-level safety evaluator that monitors all eligible non-task objects, reaching 97.38\% aggregate Scene Safety and 59.38\% Safe Success.
Abstract:Vision-language-action (VLA) policies have demonstrated impressive capabilities in generalizable robotic manipulation, but their deployment in the real world remains challenging due to potential collisions involving different parts of the robot, manipulated objects, and the surrounding environment. Existing inference-time VLA safety frameworks typically rely on simplified end-effector-centered representations that do not explicitly model the full articulated robot and attached-object geometry. In this paper, we present WBAG, a safety framework that models the robot's whole-body and grasp-dependent attached geometry. WBAG constructs a grasp-conditioned safe set that adapts the protected geometry as objects are grasped, then converts this evolving geometry into differentiable CBF constraints that minimally modify the VLA's native six-dimensional operational-space action for collision avoidance across robot, scene, and attached geometry. On the SafeLIBERO benchmark, a variant of LIBERO augmented with obstacles for safety evaluation, WBAG achieves the best overall safety and safe task success among the evaluated methods under a scene-level safety evaluator that monitors all eligible non-task objects, reaching 97.38\% aggregate Scene Safety and 59.38\% Safe Success.
Robotics (cs.RO)
arXiv:2610.01083 [cs.RO]
(or arXiv:2610.01083v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.01083 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.
Full text under the publisher's open licence · 260 words
Related coverage
Links #5: 2026/07
LessWrong · Jul 31, 2026
EgoExplorer: A Vision-Motion-Touch Observatory for Human Loco-Manipulation In Situ
arXiv · Oct 14, 2026
The Entangled Dimensions of Decision Theory
LessWrong · Jul 30, 2026
RoboReact: Agentic Skill Distillation from Generated Egocentric Videos for Generalizable Whole-Body Manipulation
arXiv · Aug 5, 2026