Airbot_MMK2_storage_bowl

该数据集使用基于LeRobot的扩展格式，并完全兼容LeRobot。它包含46个总片段，10955帧，帧率为30FPS，数据集大小为432.12 MB。机器人名称为Airbot_MMK2，末端执行器类型为五指夹爪。数据集包含四个摄像头视图：cam_head_rgb、cam_left_wrist_rgb、cam_right_wrist_rgb和cam_front_rgb。场景为餐饮->餐厅，对象包括碗（未知）和盆（未知）。任务描述为用手拿起桌子上的碗并将其放入黄色的盆中。数据集还包括6个不同的子任务和丰富的注释以支持多样化的学习方法。

This dataset adopts an extended format based on LeRobot and is fully compatible with LeRobot. It contains 46 total segments and 10,955 frames, with a frame rate of 30 FPS, and has a total size of 432.12 MB. The robot is named Airbot_MMK2, equipped with a five-finger gripper as its end-effector. The dataset includes four camera views: cam_head_rgb, cam_left_wrist_rgb, cam_right_wrist_rgb, and cam_front_rgb. The scenario is categorized as catering → restaurant, with the involved objects including bowls (unknown) and basins (unknown). The task is to pick up the bowl on the table and place it into the yellow basin. Furthermore, the dataset includes six distinct subtasks and comprehensive annotations to support diverse learning methodologies.

原始信息汇总

Airbot_MMK2_storage_bowl 数据集概述

数据集基本信息

• 数据集名称：Airbot_MMK2_storage_bowl
• 任务类别：机器人学
• 语言：英语
• 许可协议：Apache-2.0
• 格式兼容性：基于LeRobot的扩展格式，完全兼容LeRobot。

数据集规模与统计

指标	数值
总情节数	46
总帧数	10955
总任务数	6
总视频数	184
总数据块数	1
数据块大小	1000
帧率	30 FPS
状态维度	36
动作维度	36
相机视角数	4
数据集大小	432.12 MB

任务与场景

• 主要任务指令：pick the bowls on the table by hands and place them into the yellow basin.
• 标准化任务描述：pick the bowls on the table by hands and place them into the yellow basin.
• 场景类型：catering->restaurant
• 操作对象：
  • bowl(unknown)
  • basin(unknown)
• 原子动作：grasp, pick, place
• 子任务：
  1. Grasp the pink bowl with the left gripper (索引: 0)
  2. End (索引: 1)
  3. Grasp the white bowl with the right gripper (索引: 2)
  4. Place the pink bowl into the yellow storage box with the left gripper (索引: 3)
  5. Place the white bowl into the yellow storage box with the right gripper (索引: 4)
  6. null (索引: 5)

机器人配置

• 机器人名称：Airbot_MMK2
• 代码库版本：v2.1
• 末端执行器类型：five_finger_gripper
• 遥操作类型：Due to some reasons, this dataset temporarily cannot provide the teleoperation type information.
• 操作类型：Due to some reasons, this dataset temporarily cannot provide the operation type information.
• 环境类型：Due to some reasons, this dataset temporarily cannot provide the environment type information.

硬件与传感器

• 传感器列表：
  • cam_head_rgb
  • cam_left_wrist_rgb
  • cam_right_wrist_rgb
  • cam_front_rgb
• 相机信息：
  • cam_head_rgb: dtype=video, shape=480x640x3, resolution=640x480, codec=av1, pix_fmt=yuv420p
  • cam_left_wrist_rgb: dtype=video, shape=480x640x3, resolution=640x480, codec=av1, pix_fmt=yuv420p
  • cam_right_wrist_rgb: dtype=video, shape=480x640x3, resolution=640x480, codec=av1, pix_fmt=yuv420p
  • cam_front_rgb: dtype=video, shape=480x640x3, resolution=640x480, codec=av1, pix_fmt=yuv420p
• 坐标系定义：right-hand-frame
• 维度与单位：
  • 关节旋转：radian
  • 末端执行器旋转：end_rotation_dim
  • 末端执行器平移：end_translation_dim

数据结构与组织

• 数据拆分：训练集包含情节 0 至 45。

• 数据文件组织模式：

  • 数据路径模式：data/chunk-{id}/episode_{id}.parquet
  • 视频路径模式：videos/chunk-{id}/observation.images.cam_left_wrist_rgb/episode_{id}.mp{id}
  • 分块：数据被组织成 1 个数据块，每个块大小为 1000。

• 目录结构：

  Airbot_MMK2_storage_bowl_qced_hardlink/ |-- annotations | |-- eef_acc_mag_annotation.jsonl | |-- eef_direction_annotation.jsonl | |-- eef_velocity_annotation.jsonl | |-- gripper_activity_annotation.jsonl | |-- gripper_mode_annotation.jsonl | |-- scene_annotations.jsonl | -- subtask_annotations.jsonl |-- data | -- chunk-000 | |-- episode_000000.parquet | |-- episode_000001.parquet | |-- episode_000002.parquet | |-- episode_000003.parquet | |-- episode_000004.parquet | |-- episode_000005.parquet | |-- episode_000006.parquet | |-- episode_000007.parquet | |-- episode_000008.parquet | |-- episode_000009.parquet | |-- episode_000010.parquet | -- episode_000011.parquet | -- ... (34 more entries) |-- meta | |-- episodes.jsonl | |-- episodes_stats.jsonl | |-- info.json | -- tasks.jsonl -- videos -- chunk-000 |-- observation.images.cam_front_rgb |-- observation.images.cam_head_rgb |-- observation.images.cam_left_wrist_rgb -- observation.images.cam_right_wrist_rgb

• 可用标注：

  • eef_acc_mag_annotation.jsonl
  • eef_direction_annotation.jsonl
  • eef_velocity_annotation.jsonl
  • gripper_activity_annotation.jsonl
  • gripper_mode_annotation.jsonl
  • scene_annotations.jsonl
  • subtask_annotations.jsonl

特征与数据模式

数据集包含以下主要特征（基于提供的YAML描述）：

• observation.images.cam_head_rgb：视频格式，形状 480x640x3。
• observation.images.cam_left_wrist_rgb：视频格式，形状 480x640x3。
• observation.images.cam_right_wrist_rgb：视频格式，形状 480x640x3。
• observation.images.cam_front_rgb：视频格式，形状 480x640x3。
• observation.state：float32类型，形状为36，包含左右臂及左右手各关节的角度（弧度）。
• action：float32类型，形状为36，结构与状态观测对应。
• 其他特征包括时间戳、帧索引、情节索引、任务索引、子任务标注、场景标注以及多种末端执行器状态和动作数据。

作者与来源

• 贡献者：北京智源人工智能研究院的RoboCOIN团队。
• 数据集标签：RoboCOIN, LeRobot
• 相关链接：
  • 主页：https://flagopen.github.io/RoboCOIN/
  • 论文：https://arxiv.org/abs/2511.17441
  • 代码仓库：https://github.com/FlagOpen/RoboCOIN

使用条款与引用

• 访问条件：使用此数据集需同意在研究/出版物中引用相关论文（详见“引用”部分），并同意不将该数据集用于对人类受试者造成伤害的实验。

• 引用格式： bibtex @article{robocoin, title={RoboCOIN: An Open-Sourced Bimanual Robotic Data Collection for Integrated Manipulation}, author={Shihan Wu, Xuecheng Liu, Shaoxuan Xie, Pengwei Wang, Xinghang Li, Bowen Yang, Zhe Li, Kai Zhu, Hongyu Wu, Yiheng Liu, Zhaoye Long, Yue Wang, Chong Liu, Dihan Wang, Ziqiang Ni, Xiang Yang, You Liu, Ruoxuan Feng, Runtian Xu, Lei Zhang, Denghang Huang, Chenghao Jin, Anlan Yin, Xinlong Wang, Zhenguo Sun, Junkai Zhao, Mengfei Du, Mingyu Cao, Xiansheng Chen, Hongyang Cheng, Xiaojie Zhang, Yankai Fu, Ning Chen, Cheng Chi, Sixiang Chen, Huaihai Lyu, Xiaoshuai Hao, Yequan Wang, Bo Lei, Dong Liu, Xi Yang, Yance Jiao, Tengfei Pan, Yunyan Zhang, Songjing Wang, Ziqian Zhang, Xu Liu, Ji Zhang, Caowei Meng, Zhizheng Zhang, Jiyang Gao, Song Wang, Xiaokun Leng, Zhiqiang Xie, Zhenzhen Zhou, Peng Huang, Wu Yang, Yandong Guo, Yichao Zhu, Suibing Zheng, Hao Cheng, Xinmin Ding, Yang Yue, Huanqian Wang, Chi Chen, Jingrui Pang, YuXi Qian, Haoran Geng, Lianli Gao, Haiyuan Li, Bin Fang, Gao Huang, Yaodong Yang, Hao Dong, He Wang, Hang Zhao, Yadong Mu, Di Hu, Hao Zhao, Tiejun Huang, Shanghang Zhang, Yonghua Lin, Zhongyuan Wang and Guocai Yao}, journal={arXiv preprint arXiv:2511.17441}, url = {https://arxiv.org/abs/2511.17441}, year={2025}, }

• 附加引用建议：LeRobot框架：https://github.com/huggingface/lerobot