CarryOn：用于个性化且可靠机器人辅助的双循环控制装置
CarryOn: A Dual-Loop Harness for Personalized and Reliable Robot Assistance
AI总结：
提出CarryOn双循环装置增强VLA策略，实现机器人辅助的个性化与可靠执行，降低安全违规率并提升恢复成功率。
摘要：
尽管通用视觉-语言-动作（VLA）模型已实现多样的开放世界操作，但现实中的机器人辅助仍需两项其目前缺乏的关键能力：个性化与可靠执行。首先，执行“给我做个三明治”这类指令时，需将子任务分解基于持久的用户偏好，而非依赖单一通用方案。其次，可靠部署要求可解释的安全保障，以主动预防危险并从偶尔的物理故障中优雅恢复。为解决这些缺口，我们提出CarryOn，一种用于增强冻结VLA策略的双循环控制装置。规划循环维护持久的用户级与场景级上下文，该上下文持续与新观测结果建立关联，并通过交互修正，支持跨对话与变化场景的个性化任务选择及执行关联。执行循环主动监督策略 rollout，保护动作免受预测的危险接触。当执行失败时，系统调用可复用的基于控制的恢复技能，以恢复任务进度，并通过对早期失败的反思进行优化。在真实机器人与仿真环境中，CarryOn 针对不同用户与场景分发符合偏好的子任务，将已进行安全微调的VLA的违规率从52.6%降至6.1%，并在复用优化后的技能时，将恢复成功率从13%提升至80%。
英文摘要：
While generalist vision-language-action (VLA) models have enabled diverse open-world manipulation, real-world robotic assistance requires two critical capabilities they currently lack: personalization and reliable execution. First, executing an instruction like "make me a sandwich" requires grounding subtask decomposition in persistent user preferences rather than relying on a single generic solution. Second, reliable deployment demands explainable safety assurance that proactively prevents hazards and gracefully recovers from occasional physical failures. To address these gaps, we present CarryOn, a dual-loop harness designed to augment a frozen VLA policy. The planning loop maintains persistent user-level and scene-level context that is continually grounded against new observations and revised through interaction, enabling personalized task selection and execution grounding across conversations and changing scenes. The execution loop actively supervises policy rollouts, shielding actions from predicted hazardous contact. When execution fails, the system invokes reusable control-based recovery skills that restore task progress and are refined through reflection on earlier failures. On a real robot and in simulation, CarryOn dispatches preference-consistent subtasks across users and scenes, reduces the violation rate of an already safety-finetuned VLA from 52.6% to 6.1%, and increases recovery success from 13% to 80% when refined skills are reused.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
