# HubEgo-100K发布：万小时第一人称技能视频数据集助力机器人泛化

> 原文链接：https://robotics.hellobike.com/papers/2026-11-18-2611-21938-980f7d65
> 频道：papers | 更新日期：2026-11-18 | 标签：auto, arxiv
> 核心摘要：针对模仿学习泛化瓶颈，HubEgo-100K v1.0开源10124小时1080p第一人称视频及100Hz IMU数据。覆盖15548个真实工作场景与30类工种，83%为职场环境，提供LeRobot格式标注，验证了环境多样性对策略性能的关键作用。

# HubEgo-100K发布：万小时第一人称技能视频数据集助力机器人泛化

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[HubEgo-100K: In-the-Wild Egocentric Video of Skilled Work for Robot Learning](https://arxiv.org/abs/2611.21938)（arXiv 2611.21938）

## 核心内容

针对机器人模仿学习在训练分布外泛化能力差的问题，研究指出策略泛化更依赖环境与物体多样性而非单一场景演示数量。为此发布的HubEgo-100K v1.0数据集包含10124小时1080p第一人称视频及同步100Hz IMU数据，由12227名贡献者在15548个独立环境中采集。数据涵盖11种设置类型、30个工作类别及56种任务类型，其中83%时长来自真实职场而非日常生活，旨在通过大规模异构数据解决技能迁移难题。

该数据集包含72754个片段，提供机器生成的动作与语言标注，采用`LeRobot`和`MCAP`格式并在CC BY 4.0协议下开源。数据采集使用头戴手机摄像头，确保了手部可见性与惯性传感数据的对齐，这两项指标已被证实与策略性能强相关。研究显示在随机排序下新贡献者与词汇持续涌现，证明了数据的高熵特性。目前v1.0版本已提供完整数据表与隐私处理文档，后续版本将扩展至10万小时并增加基准测试。

## 关键看点

- **数据规模与模态**：v1.0释放10124小时1080p视频及同步100Hz IMU数据，包含72754个片段，支持多模态具身策略训练。
- **场景多样性**：覆盖15548个独立环境、30类工作类别及56种任务，83%数据源自真实职场，显著区别于日常活动数据集。
- **工程兼容性**：采用`LeRobot`与`MCAP`标准格式存储，附带机器生成动作/语言标注及CC BY 4.0许可，降低接入成本。
- **关键特征验证**：量化评估手部可见性与惯性传感等6项属性，证实数据多样性随样本量持续增长，符合泛化 scaling law。

## 工程与产业观察

该数据集填补了具身智能领域大规模真实职场技能数据的空白，验证了“环境多样性优于单场景重复”的scaling规律。其提供的同步IMU与手部视角数据，直接适配当前主流的视觉-动作联合建模需求。对于算法工程师而言，这是验证Sim-to-Real或跨域泛化算法的高价值基准；对产业界，则提示了利用低成本人类视频预训练基础模型以解决长尾技能获取的工程路径。

## 英文摘要 (Abstract)

Robot policies trained by imitation generalize poorly outside the environments and objects seen in training. Published data-scaling studies find that generalization grows with the diversity of training environments and objects rather than with more demonstrations per environment. Robot data is costly to collect across many environments, whereas egocentric human video can be recorded in the wild at scale; existing egocentric video, however, mostly shows daily-life activity and contains little skilled work in real workplaces. We present HubEgo-100K, a dataset built for diversity. This version of the paper covers the first release, v1.0: 10,124 hours of 1080p egocentric video with synchronized 100 Hz inertial (IMU) data from one head-mounted phone camera. In v1.0, 12,227 contributors record their own unscripted work in 15,548 environments (each one contributor's own setting of one type, such as their kitchen or workshop) across 11 setting types, 30 work categories, 56 task types and 35 action families, with 83% of hours in workplace settings. The release comprises 72,754 episodes with machine-generated action and language annotations, in LeRobot and MCAP formats under CC BY 4.0. We measure diversity along each axis, show that inside v1.0, with episodes in random order, new contributors and words keep appearing, and relate the measurements to published evidence on six data properties tied to policy performance, including hand visibility and inertial sensing. We also document collection, curation, privacy measures and data format, and provide a datasheet. Later versions of this paper will add benchmarks and cover further releases toward 100,000 hours.

作者：Javad Amirian, Armin Kiani, Lo\"ic Schaeffer, Tim Sprecher, Mohamed Taha, Karthik Yadati
:::

## 原始参考来源 (Sources)

- [arXiv 2611.21938：HubEgo-100K: In-the-Wild Egocentric Video of Skilled Work for Robot Learning](https://arxiv.org/abs/2611.21938)