CarryOn: dual-loop harness adds personalization and safe recovery to frozen VLA robots
Zixuan Wang · hf · 2026-10-16
A HKUST/CUHK/SmartMore team introduced CarryOn, a dual-loop harness that augments frozen vision-language-action (VLA) policies for real-world robot assistance, with open-source code.
- Planning loop: maintains persistent user- and scene-level context, grounded against new observations and revised via interaction, enabling personalized task decomposition across conversations and scenes.
- Execution loop: actively supervises policy rollouts, shielding actions from predicted hazards and invoking reusable control-based recovery skills refined through reflection when execution fails.
- Validated on a real robot and in simulation: dispatches preference-consistent subtasks and reduces violation rates of an already safety-tuned policy.
More from Embodied
- DynaRobotics CEO Lindon Gao on running deployment and research feedback loops in parallel — micoolcho · 2026-10-16
- Baidu to give 1,000+ Hong Kong residents free robotaxi rides in city-first public trial — Baidu_Inc · 2026-10-16
- Auki demos spatial harness: robots and devices share spatial context via AR — broodsugar · 2026-10-16
- Tesla's 2-DOF car becomes 6-DOF flying demo — easier control problem than humanoids — micoolcho · 2026-10-16
- XREAL Aura, Android XR glasses powered by Gemini, named a TIME Best Invention of 2026 — DuRuofei · 2026-10-16
- AutoAgent0: VLM-powered agentic runtime lifts long-tail driving success by 28.3pp — yining_hong · 2026-10-16