HiFi-UMI：僅從高保真UMI資料學習可部署的操作策略

摘要

學習可部署的操作策略面臨瓶頸，原因在於同時具備高保真度與可擴展性的資料極為稀缺。真實機器人遙控操作精準但擴展成本高昂；無機器人UMI資料收集則易於擴展，現行實務主要將所得資料用於預訓練，並在後訓練階段加入少量真實機器人資料作為「錨點」。我們探討：提升無機器人UMI資料的保真度（而非縮減真實機器人資料比例），是否能移除該錨點。我們提出HiFi-UMI，這是一套可攜式UMI資料生產系統，針對軌跡精度、夾爪間相對位姿、同步性與視野共同設計：頭戴式離線立體慣性SLAM、原生而非重建的相對位姿、共享微秒級GPIO觸發器，以及每隻手配備兩顆涵蓋約200度的廣角攝影機。該系統在無外部追蹤基礎設施下，可達到工作空間內末端執行器3公釐的局部精度。利用此資料庫，我們展示零機器人後訓練：僅以HiFi-UMI示範資料進行後訓練的策略，可直接部署於真實機器人，並在三個涵蓋視覺-語言-動作與世界-動作模型系列的骨幹架構上，與領域內遙控操作表現相當——在StarVLA-QwenPI、OpenPI-pi_0.5與LingBot-VA上的成功率差異分別為-2.5、+3.1與-0.6個百分點；最強的策略在一項精密插入任務上達到85%的成功率，儘管遙控操作基準線是在評估場景中收集的，而無任何HiFi-UMI軌跡來自該場景。使用同一資料庫中4,000小時的資料進行預訓練，在十項未見過任務上的動作誤差降低了41%，並在StarVLA-QwenPI上進一步將真實機器人成功率提升18.1個百分點。我們開源HiFi-UMI-2K，包含2,000小時微秒同步、超廣視野的示範資料，每段資料皆自動重建並透過模擬重演驗證，為機器人學習社群提供大規模、高保真度的資源。

English

Learning deployable manipulation policies is bottlenecked by the scarcity of data that is both high-fidelity and scalable. Real-robot teleoperation is accurate but costly to scale; robot-free UMI capture scales readily, and current practice uses the resulting data mainly for pre-training, adding a small real-robot "anchor" at post-training. We ask whether raising the fidelity of robot-free UMI data, rather than shrinking the real-robot fraction, can remove that anchor. We present HiFi-UMI, a portable UMI data-production system co-designed for trajectory accuracy, inter-gripper relative pose, synchronization, and field of view: head-mounted offline stereo-inertial SLAM, native rather than reconstructed relative pose, a shared microsecond GPIO trigger, and two wide-angle cameras per hand covering ~200 degrees. It reaches 3 mm workspace-local end-effector accuracy without external tracking infrastructure. Using this corpus, we demonstrate zero-robot post-training: a policy post-trained solely on HiFi-UMI demonstrations deploys directly on a real robot and matches in-domain teleoperation across three backbones spanning the vision-language-action and world-action-model families, with success-rate differences of -2.5, +3.1, and -0.6 percentage points on StarVLA-QwenPI, OpenPI-pi_0.5, and LingBot-VA; the strongest policy reaches 85% on a precision insertion task, even though the teleoperation baseline is collected in the evaluation scene and no HiFi-UMI trajectory is. Pre-training on 4,000 hours from the same corpus lowers action error on ten unseen tasks by 41% and, on StarVLA-QwenPI, raises real-robot success by a further 18.1 percentage points. We open-source HiFi-UMI-2K, 2,000 hours of microsecond-synchronized, ultra-wide-FoV demonstrations, each automatically reconstructed and validated through simulation replay, as a large-scale, high-fidelity resource for the robot-learning community.