HyeonseokE/SO101-DualArm-teleop_fold_towel_60epi
收藏资源简介:
该数据集是一个机器人控制数据集,使用LeRobot平台创建,专注于双机械臂折叠毛巾的任务。数据集包含60个episodes,总计39115帧,帧率为30fps。数据通过远程操作(teleop)收集,机器人类型为双机械臂跟随器(bi_so_follower)。数据集包括动作数据(如左右机械臂的关节位置和夹爪位置)、观察数据(如状态观察,包括机械臂关节位置,以及图像观察,来自三个摄像头:左腕、左顶部和右腕,分辨率均为480x640,RGB格式)。数据以parquet文件和视频文件存储,总数据大小约为100MB,视频文件大小约为200MB。数据集用于训练和评估机器人学习算法,支持多模态输入(状态和视觉)。
This dataset is a robotics control dataset created using the LeRobot platform, focusing on the task of folding a towel with dual arms. It contains 60 episodes, totaling 39,115 frames at a frame rate of 30 fps. The data was collected via teleoperation using a dual-arm follower robot (bi_so_follower). The dataset includes action data (e.g., joint positions and gripper positions for both left and right arms) and observation data (e.g., state observations comprising joint positions, and image observations from three cameras: left wrist, left top, and right wrist, all with a resolution of 480x640 and RGB format). Data is stored in parquet files and video files, with a total data size of approximately 100 MB and video file size of about 200 MB. The dataset is intended for training and evaluating robot learning algorithms, supporting multimodal inputs (state and vision).




