sobit_home-pickup_the_block-left-rel-gazebo
收藏Hugging Face2026-06-08 更新2026-06-05 收录
下载链接:
https://huggingface.co/datasets/k-valentin/sobit_home-pickup_the_block-left-rel-gazebo
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个机器人数据集,专门用于移动操作器(mobile_imulator)的机器人学习任务。数据集包含202个剧集(episodes),总计19483帧(frames),专注于一个单一任务。数据以parquet格式存储,视频以mp4格式存储,帧率为6fps。数据集的特征包括:动作(action),维度为19,涵盖头部、身体、左臂、左腕和基座等关节的控制;观测状态(observation.state),维度为19,表示关节状态;图像观测来自头部相机(分辨率480x640)和左手相机(分辨率1200x1920),提供RGB视频数据。此外,还包括时间戳、帧索引、剧集索引、任务索引等元数据。机器人信息显示,机器人名为sobit_home,类型为移动操作器,具有头部、身体、左右臂、左右末端执行器和移动基座等可动部分,并配备RGB相机传感器。数据集由Keith Valentin在Soka University, Japan创建,使用Apache 2.0许可证。
This dataset is a robotics dataset specifically designed for robot learning tasks involving mobile manipulators (mobile_imulator). It contains 202 episodes, totaling 19,483 frames, and focuses on a single task. The data is stored in parquet format, with videos in mp4 format at a frame rate of 6fps. The dataset features include: action (dimension 19), covering control of joints such as the head, body, left arm, left wrist, and base; observation.state (dimension 19), representing joint states; image observations from a head camera (resolution 480x640) and a left-hand camera (resolution 1200x1920), providing RGB video data. Additionally, metadata such as timestamps, frame index, episode index, and task index are included. The robot information indicates the robot is named sobit_home, of type mobile manipulator, with movable parts including head, body, left and right arms, left and right end-effectors, and a mobile base, equipped with RGB camera sensors. The dataset was created by Keith Valentin at Soka University, Japan, and uses the Apache 2.0 license.
提供机构:
k-valentin创建时间:
2026-06-05
原始信息汇总
数据集概述
- 数据集名称: sobit_home-pickup_the_block-left-rel-gazebo
- 许可证: Apache-2.0
- 任务类别: 机器人学 (Robotics)
- 标签: LeRobot
数据集基本信息
| 项目 | 值 |
|---|---|
| 代码版本 | v3.0 |
| 机器人类型 | 移动操作手 (mobile_manipulator) |
| 总片段数 | 202 |
| 总帧数 | 19483 |
| 总任务数 | 1 |
| 块大小 | 1000 |
| 数据文件大小 | 100 MB |
| 视频文件大小 | 200 MB |
| 帧率 | 6 FPS |
| 训练/测试划分 | 训练: 0:202 (全部) |
数据结构
- 数据路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
特征(Features)
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
action |
float32 | [19] | 19维动作指令(头部、身体、左臂、左手手指关节 + 移动基座) |
observation.state |
float32 | [19] | 19维状态观测(与动作维度一致) |
observation.images.head_camera |
video | [3, 480, 640] | 头部摄像头视频,分辨率 640x480,H.264 编码,6 FPS |
observation.images.hand_left_camera |
video | [3, 1200, 1920] | 左手摄像头视频,分辨率 1920x1200,H.264 编码,6 FPS |
timestamp |
float32 | [1] | 时间戳 |
frame_index |
int64 | [1] | 帧索引 |
episode_index |
int64 | [1] | 片段索引 |
index |
int64 | [1] | 全局索引 |
task_index |
int64 | [1] | 任务索引 |
动作与状态维度名称
- 头部:
head_pan_joint,head_tilt_joint - 身体:
body_lift_joint - 左臂:
arm_left_shoulder_tilt_joint,arm_left_upper_roll_joint,arm_left_upper_flex_joint,arm_left_elbow_joint,arm_left_wrist_tilt_joint,arm_left_wrist_roll_joint - 左手手指:
hand_left_finger_l_mcp_joint至hand_left_finger_r_dip_joint(共6个) - 移动基座:
base_x,base_y,base_theta
机器人信息
- 名称: sobit_home
- 版本: 1.0.0
- 形态: 移动操作手 (mobile_manipulator)
- 可动部件:
- 头部 (head): 2个关节
- 身体 (body): 1个关节
- 左臂 (arm_left): 6个关节
- 右臂 (arm_right): 6个关节
- 左端执行器 (end_effect_left): 8个关节
- 右端执行器 (end_effect_right): 8个关节
- 移动基座 (mobile_base): 8个关节
- 传感器: RGB摄像头
- 头部摄像头: 640x480
- 左手摄像头: 1920x1200
- 右手摄像头: 1920x1200
用户信息
- 姓名: Keith Valentin
- 邮箱: e25d5303@soka-u.jp
- 位置: 日本创价大学
引用信息
目前暂无可用的 BibTeX 引用信息。



