analiza aiqbench
This research addresses the critical safety gap in deploying Vision-Language-Action models for physical manipulation by introducing a passivity-shielded compliance prior that enforces energy constraints at high frequencies. By decoupling semantic reasoning from low-level control authority, the framework ensures stable performance in contact-rich environments like connector insertion without relying on direct motor commands from foundation models. The work represents a significant step toward safely integrating large AI models into real-world robotic systems where force dynamics must be rigorously regulated.
PaCo-VLA: Passivity-Shielded Compliance Prior for Contact-Rich Vision-Language-Action Manipulation
arXiv:2606.00515v2 Announce Type: replace-cross Abstract: Contact-rich manipulation demands both high-level semantic reasoning and the safe regulation of high-frequency contact dynamics. While Vision-Language-Action (VLA) models provide unprecedented semantic generalization, their low-rate outputs lack the reliability required for direct plant authority in force-sensitive tasks. To bridge this semantic-to-control gap, we introduce PaCo-VLA, a passivity-shielded compliance prior that recasts the VLA interface. Rather than trusting VLAs with direct motor commands, PaCo-VLA treats network outputs as task-level compliance proposals: semantic bindings, task stages, and admittance schedules. A high-frequency, proposal-independent passivity shield governs these proposals through energy-tank accounting and boundary checks, preventing invalid, stale, or unverified model predictions from bypassing low-level contact physics. This decoupled architecture also enables causal evaluation, isolating semantic contributions from geometric shortcuts. Extensive simulated and real-world connector-insertion experiments demonstrate that PaCo-VLA achieves superior precision over unshielded VLA baselines, sustaining zero passivity violations even under adversarial compliance shifts. This framework establishes a provably sampled-passive runtime contract at the admittance port and provides a runtime interface for deploying foundation models in contact-rich domains.
Abstract:Contact-rich manipulation demands both high-level semantic reasoning and the safe regulation of high-frequency contact dynamics. While Vision-Language-Action (VLA) models provide unprecedented semantic generalization, their low-rate outputs lack the reliability required for direct plant authority in force-sensitive tasks. To bridge this semantic-to-control gap, we introduce PaCo-VLA, a passivity-shielded compliance prior that recasts the VLA interface. Rather than trusting VLAs with direct motor commands, PaCo-VLA treats network outputs as task-level compliance proposals: semantic bindings, task stages, and admittance schedules. A high-frequency, proposal-independent passivity shield governs these proposals through energy-tank accounting and boundary checks, preventing invalid, stale, or unverified model predictions from bypassing low-level contact physics. This decoupled architecture also enables causal evaluation, isolating semantic contributions from geometric shortcuts. Extensive simulated and real-world connector-insertion experiments demonstrate that PaCo-VLA achieves superior precision over unshielded VLA baselines, sustaining zero passivity violations even under adversarial compliance shifts. This framework establishes a provably sampled-passive runtime contract at the admittance port and provides a runtime interface for deploying foundation models in contact-rich domains.
Robotics (cs.RO); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
arXiv:2606.00515 [cs.RO]
(or arXiv:2606.00515v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.00515 Focus to learn more arXiv-issued DOI via DataCite
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 · 263 cuvinte
Subiecte conexe
Safe Learning for Contact-Rich Robot Tasks: A Survey from Classical Learning-Based Methods to Safe Foundation Models
arXiv · 22 sept. 2026
Compliance for Free: Learning Identifiable Impedance via Bilateral Teleoperation
arXiv · 18 sept. 2026
Building a Foundation Stack for General-Purpose Robots
IEEE Spectrum · 13 iul. 2026
PHASE: Compliance-Enabled Tactile Phase Retrieval for Few-Shot Insertion Learning
arXiv · 28 sept. 2026