dataset_env_starbuckswall9vegetabledistractor_task_pickcube_lift_epi_10_ctrl_cartesian
收藏资源简介:
该数据集是一个用于机器人学习的数据集,由LeRobot工具创建,专门针对拾取立方体任务。数据集包含10个训练集,共1439帧,使用panda机器人类型。数据以parquet文件格式存储,视频文件以MP4格式提供,帧率为15fps。特征包括从主要摄像头和手腕摄像头捕获的图像观察、机器人状态(如笛卡尔坐标、关节角度、夹爪状态)、目标状态以及动作数据。所有观察和动作数据以浮点型或整型表示,适用于机器人控制算法的训练和评估。数据集结构详细,支持机器人任务中的强化学习或模仿学习应用。
This dataset is designed for robot learning, created by the LeRobot tool, specifically for the cube picking task. It contains 10 training sets with a total of 1439 frames, using the panda robot type. Data is stored in parquet file format, with video files provided in MP4 format at a frame rate of 15fps. Features include image observations captured from the main camera and wrist camera, robot states (such as Cartesian coordinates, joint angles, gripper status), target states, and action data. All observation and action data are represented as floating-point or integer types, suitable for training and evaluating robot control algorithms. The dataset has a detailed structure and supports applications in reinforcement learning or imitation learning for robotic tasks.
数据集概述
- 数据集名称: dataset_env_starbuckswall9vegetabledistractor_task_pickcube_lift_epi_10_ctrl_cartesian
- 许可证: Apache-2.0
- 任务类别: 机器人学(Robotics)
- 标签: LeRobot
- 主页: 暂无
- 论文: 暂无
数据集结构
- 机器人类型: Panda(Franka Emika Panda)
- 总片段数 (Episodes): 10
- 总帧数 (Frames): 1439
- 总任务数 (Tasks): 1
- 数据分块大小: 1000
- 数据文件大小: 约100 MB
- 视频文件大小: 约200 MB
- 帧率 (FPS): 15
- 数据集划分: 全部10个片段用于训练(train: 0:10)
数据特征 (Features)
数据集包含以下特征,存储在Parquet文件中,视频数据以MP4格式存储:
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
observation.images.primary |
视频 | (256, 256, 3) | 主摄像头图像,分辨率为256x256,RGB三通道,帧率15 FPS,编码格式av1 |
observation.images.wrist |
视频 | (256, 256, 3) | 腕部摄像头图像,分辨率为256x256,RGB三通道,帧率15 FPS,编码格式av1 |
observation.state.cartesian |
float32 | (6,) | 末端执行器笛卡尔状态,包含x、y、z、roll、pitch、yaw |
observation.state.gripper |
float32 | (1,) | 夹爪状态 |
observation.state.joints |
float32 | (7,) | 7个关节位置状态 |
observation.state.target |
float32 | (6,) | 目标笛卡尔状态,包含target_x、target_y、target_z、target_roll、target_pitch、target_yaw |
observation.state |
float32 | (20,) | 组合状态,包含上述所有状态(cartesian、gripper、joints、target) |
action |
float32 | (7,) | 动作指令,包含x、y、z、roll、pitch、yaw、gripper |
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 |
引用信息
暂无BibTeX引用信息。



