HyeonseokE/SO101-PickandPlace_Ours_20epi
收藏资源简介:
该数据集是使用LeRobot创建的,专注于机器人操作任务,具体为SO101-PickandPlace任务。数据集包含20个episodes,总计6586帧,数据以parquet文件格式存储。它涉及so101_follower机器人类型,数据特征包括观察状态(如6个关节位置:肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转、夹爪位置)、动作、图像观察(顶部和左腕摄像头的视频,分辨率为480x640,帧率为10fps)、末端执行器位置(以xyzrpy坐标表示)、夹爪二进制状态、技能信息(如自然语言描述、验证问题、技能类型、进度、目标关节位置、目标机器人xyzrpy位置、目标夹爪位置)、子任务信息(如自然语言描述、对象名称、目标位置)以及时间戳、帧索引、episode索引等元数据。数据集适用于机器人学习、模仿学习或强化学习研究,旨在支持机器人抓取和放置任务的开发与评估。
This dataset was created using LeRobot and focuses on robotics manipulation tasks, specifically the SO101-PickandPlace task. It contains 20 episodes with a total of 6586 frames, stored in parquet file format. The dataset involves the so101_follower robot type, and its features include observation states (e.g., 6 joint positions: shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, gripper), actions, image observations (videos from top and left wrist cameras with a resolution of 480x640 and a frame rate of 10fps), end-effector positions (in xyzrpy coordinates), gripper binary state, skill information (such as natural language descriptions, verification questions, skill type, progress, goal joint positions, goal robot xyzrpy positions, goal gripper positions), subtask information (e.g., natural language descriptions, object names, target positions), and metadata like timestamps, frame indices, and episode indices. The dataset is suitable for robotics learning, imitation learning, or reinforcement learning research, aiming to support the development and evaluation of robot grasping and placement tasks.




