遇见数据集

multiple_box_experiment_20260527_184122

收藏
Hugging Face2026-05-27 更新2026-05-27 收录
官方服务:

资源简介:

该数据集是由LeRobot创建的机器人操作数据集,用于记录机器人执行任务时的状态和动作信息。数据集包含一个episode,共630帧,采样率为30fps。特征包括动作(6维浮点数组,对应肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪的位置)、观测状态(117维浮点数组,涵盖关节位置、原始电流、原始负载、硬件速度、速度、原始扭矩、签名电流、签名扭矩、滤波速度、估计外部扭矩、估计扭矩方差以及模型计算的电机扭矩、重力扭矩、摩擦扭矩、科里奥利扭矩、惯性扭矩和模型扭矩等详细信息,还包括工具中心点(TCP)的力和力矩以及HPI夹爪的外部扭矩、位置和速度)、两个图像观测(顶部和角落视角的视频,分辨率480x640,3通道,编码为av1),以及时间戳、帧索引、episode索引、索引和任务索引。数据以parquet格式存储,视频以mp4格式存储,机器人类型为so_follower,适用于机器人学习和控制任务。

This dataset is a robot operation dataset created by LeRobot, designed to record the state and action information during robot task execution. It contains one episode with 630 frames at a sampling rate of 30fps. Features include actions (a 6-dimensional float array corresponding to shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper position), observation states (a 117-dimensional float array covering joint positions, raw current, raw load, hardware speed, velocity, raw torque, signed current, signed torque, filtered speed, estimated external torque, estimated torque variance, as well as model-calculated motor torque, gravity torque, friction torque, Coriolis torque, inertial torque, and model torque, along with force and torque at the Tool Center Point (TCP) and external torque, position, and speed of the HPI gripper), two image observations (videos from top and corner perspectives, with a resolution of 480x640, 3 channels, encoded as av1), as well as timestamps, frame index, episode index, index, and task index. Data is stored in parquet format, videos in mp4 format, and the robot type is so_follower, suitable for robot learning and control tasks.

提供机构:
ecappiell
创建时间:
2026-05-27
原始信息汇总

数据集概述

  • 数据集名称: multiple_box_experiment_20260527_184122
  • 许可证: Apache-2.0
  • 任务类别: 机器人(Robotics)
  • 标签: LeRobot
  • 创建工具: 基于 LeRobot 创建

数据集结构

  • 配置: 默认配置(default),数据文件为 data/*/*.parquet
  • 数据格式: 包含元信息文件 meta/info.json,数据以 Parquet 和视频文件形式组织
  • 帧率(FPS): 30
  • 总集数(Episodes): 1
  • 总帧数(Frames): 630
  • 总任务数(Tasks): 1
  • 数据分块: 每块大小为 1000
  • 数据文件大小: 约 100 MB
  • 视频文件大小: 约 200 MB
  • 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
  • 机器人类型: so_follower
  • 数据划分: 仅包含训练集(train: 0:1

特征详情

特征名 数据类型 形状 说明
action float32 [6] 6个动作维度:肩关节、肘关节、腕关节、夹爪等位置信息
observation.state float32 [117] 117维观测状态,包含各关节位置、电流、负载、速度、扭矩、滤波/估计/模型参数、TCP力/力矩、夹爪状态等
observation.images.top video [480, 640, 3] 顶部摄像机图像,分辨率480x640,RGB三通道,AV1编码,30FPS
observation.images.corner video [480, 640, 3] 角落摄像机图像,分辨率480x640,RGB三通道,AV1编码,30FPS
timestamp float32 [1] 时间戳
frame_index int64 [1] 帧索引
episode_index int64 [1] 集数索引
index int64 [1] 全局索引
task_index int64 [1] 任务索引

引用信息

目前无可用引用信息。

二维码
社区交流群
二维码
科研交流群
商业服务