planar_3joint
收藏资源简介:
Planar_3Joint数据集是一个机器人学领域的数据集,使用LeRobot框架创建,专为机器人控制任务设计,特别适用于基于视觉的模仿学习或强化学习研究。数据集包含2000个独立的交互片段,总计458,714个数据帧,采样频率为30帧/秒。每个数据样本包含多模态观测信息:两种格式的RGB视觉输入(224x224分辨率,3通道),分别经过letterbox和stretch处理;4维的机器人关节状态(包括3个关节和1个夹爪);4维的动作指令;8维的环境状态(包含末端执行器的x和z坐标)。此外,数据集还提供了时间戳、帧索引、片段索引等元数据。数据以Parquet格式存储,总数据文件大小为100MB,相关视频文件大小为200MB。该数据集适用于训练机器人策略模型,特别是需要结合视觉感知和状态信息的控制任务。
The Planar_3Joint dataset is a robotics dataset created using the LeRobot framework. It is specifically designed for robotic control tasks and is particularly suitable for vision-based imitation learning or reinforcement learning research. The dataset contains 2000 independent interaction episodes, with a total of 458,714 data frames and a sampling frequency of 30 frames per second. Each data sample includes multimodal observation information: two formats of RGB visual inputs (224×224 resolution, 3 channels) processed via letterbox and stretch respectively; 4-dimensional robotic joint states including 3 joints and 1 gripper; 4-dimensional action commands; and 8-dimensional environmental states including the x and z coordinates of the end-effector. Additionally, the dataset provides metadata such as timestamps, frame indices, and episode indices. The data is stored in Parquet format, with a total data file size of 100 MB and associated video files totaling 200 MB. This dataset is applicable for training robotic policy models, especially for control tasks that require integrating visual perception and state information.
数据集概述
- 数据集名称:
planar_3joint - 许可证: Apache-2.0
- 任务类别: 机器人学
- 标签: LeRobot
- 创建工具: LeRobot
数据集结构与特征
- 帧率: 30 FPS
- 总片段数: 2000
- 总帧数: 458714
- 总任务数: 1
- 数据集分割: 仅包含训练集(0:2000)
数据特征
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
observation.images.base_rgb_letterbox |
image | (3, 224, 224) | 基础RGB信箱图像 |
observation.images.base_rgb_stretch |
image | (3, 224, 224) | 基础RGB拉伸图像 |
observation.state |
float32 | (4,) | 关节状态:joint_1, joint_2, joint_3, gripper |
action |
float32 | (4,) | 动作:joint_1, joint_2, joint_3, gripper |
observation.environment_state |
float32 | (8,) | 环境状态:env_0~5, ee_x, ee_z |
timestamp |
float32 | (1,) | 时间戳 |
frame_index |
int64 | (1,) | 帧索引 |
episode_index |
int64 | (1,) | 片段索引 |
index |
int64 | (1,) | 索引 |
task_index |
int64 | (1,) | 任务索引 |
文件存储
- 数据文件:
data/*/*.parquet,分块存储,路径格式data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4 - 数据大小: 数据文件约100MB,视频文件约200MB
- 分块大小: 每块1000帧
机器人类型
- 机器人类型:
lerobot_splatsim
可视化
该数据集可通过 可视化空间 进行交互式查看。




