openarm_teleop_udp_d435_merged_20260725_lerobot

这是一个使用LeRobot创建的双机械臂机器人遥操作数据集，主要用于机器人学任务。数据集包含289个episodes，总计220,424帧，以30帧每秒的速率采集。特征包括：观测状态（16维浮点数组，表示左右机械臂各7个关节的位置和夹爪位置）、动作（16维浮点数组，结构与观测状态相同）、观测速度（16维浮点数组，表示关节和夹爪速度）、观测扭矩（16维浮点数组，表示关节和夹爪扭矩），以及来自三个视角的视频观测（左腕、前视、右腕摄像头，分辨率480x640，3通道RGB视频）。此外，还包含时间戳、帧索引、episode索引、任务索引等元数据。数据集采用Apache 2.0许可证，数据以Parquet格式存储，视频以MP4格式存储，适用于机器人模仿学习或强化学习研究。This is a bimanual robotic teleoperation dataset created using LeRobot, primarily designed for robotics research tasks. The dataset contains 289 episodes, totaling 220,424 frames collected at 30 frames per second. Its features include: observational state (a 16-dimensional floating-point array representing the positions of 7 joints and the gripper position for both the left and right robotic arms), actions (a 16-dimensional floating-point array with an identical structure to the observational state), observational velocity (a 16-dimensional floating-point array representing joint and gripper velocities), observational torque (a 16-dimensional floating-point array representing joint and gripper torques), as well as video observations from three camera viewpoints: left wrist, front-facing, and right wrist cameras, with a resolution of 480x640 and 3-channel RGB video. In addition, it includes metadata such as timestamps, frame indices, episode indices, and task indices. The dataset is licensed under Apache 2.0, with data stored in Parquet format and videos stored in MP4 format. It is suitable for research on robotic imitation learning or reinforcement learning.