cell1_20260518_mohamehossam_sorting20260518_192049

这是一个机器人学数据集，使用LeRobot创建，专门用于双机械臂控制任务。数据集基于starpilot_yam_gripper机器人类型，包含6个episodes，总计3180帧数据，以30fps记录。数据以parquet文件（数据文件）和mp4视频文件形式存储，总数据大小约100MB，视频文件大小约200MB。数据集特征包括：14维动作数据（控制两个机械臂的x、y、z位置、rx、ry、rz旋转和夹爪宽度），32维观测状态数据（包括编码器角度、IMU加速度和角速度、机械臂姿态四元数、夹爪开合状态和距离），以及来自多个摄像头的观测图像：左腕0号摄像头RGB图像和深度图像、右腕0号摄像头RGB图像和深度图像、基座0号和1号摄像头RGB图像（分辨率分别为480x640和768x1024）。所有图像以视频格式存储，编码为h264。数据集仅包含训练分割（episodes 0到6），适用于机器人学习、模仿学习或强化学习任务。This is a robotics dataset developed with LeRobot, exclusively tailored for dual-arm robot control tasks. The dataset is based on the starpilot_yam_gripper robot platform, containing 6 episodes with a total of 3180 frames recorded at 30 fps. The data is stored in parquet files (raw data files) and MP4 video files, with a total raw data size of approximately 100 MB and video files totaling around 200 MB. The dataset features the following contents: 14-dimensional action data (controlling the x, y, z positions, rx, ry, rz orientations and gripper width of the two robotic arms), 32-dimensional observation state data (including encoder angles, IMU acceleration and angular velocity, robotic arm attitude quaternions, gripper opening/closing state and distance), as well as observation images from multiple cameras: RGB and depth images from the left wrist camera 0, RGB and depth images from the right wrist camera 0, and RGB images from the base camera 0 and 1 (with resolutions of 480×640 and 768×1024 respectively). All images are stored in video format and encoded with H.264. The dataset only includes the training split (episodes 0 to 6), which is applicable to robotic learning, imitation learning or reinforcement learning tasks.

原始信息汇总

数据集概述

• 名称: cell1_20260518_mohamehossam_sorting20260518_192049
• 许可证: Apache-2.0
• 任务类别: 机器人学 (robotics)
• 创建工具: LeRobot
• 机器人类型: starpilot_yam_gripper

数据规模

• 总片段数 (Episodes): 6
• 总帧数 (Frames): 3180
• 总任务数: 1
• 帧率 (FPS): 30
• 数据文件大小: 100 MB
• 视频文件大小: 200 MB
• 数据分块: 每块1000帧 (chunks_size: 1000)
• 数据划分: 训练集 (train) 包含全部6个片段 (0:6)

数据格式与结构

数据以 Parquet 和 MP4 格式存储，路径规则如下：

• 数据路径: data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet
• 视频路径: videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4

特征 (Features)

动作 (action)

• 数据类型: float32
• 维度: [14]
• 字段说明:
  • arm1_x, arm1_y, arm1_z, arm1_rx, arm1_ry, arm1_rz, arm1_grip_width
  • arm2_x, arm2_y, arm2_z, arm2_rx, arm2_ry, arm2_rz, arm2_grip_width

观测状态 (observation.state)

• 数据类型: float32
• 维度: [32]
• 字段说明: 包含双臂的编码器角度、IMU数据、位姿、夹爪状态等，具体包括：
  • 每臂的 enc_deg, imu_ax, imu_ay, imu_az, imu_gx, imu_gy, imu_gz, pose_x, pose_y, pose_z, pose_qx, pose_qy, pose_qz, pose_qw, gripper_open, gripper_distance_m

观测图像 (observation.images)

• 视频格式: h264, yuv420p
• 帧率: 30 FPS

图像名称	分辨率	说明
left_wrist_0_camera_rgb_image	480×640×3	左腕RGB相机
left_wrist_0_camera_rgb_image_depth	480×640×3	左腕RGB深度图
right_wrist_0_camera_rgb_image	480×640×3	右腕RGB相机
right_wrist_0_camera_rgb_image_depth	480×640×3	右腕RGB深度图
base_0_camera_rgb_image	480×640×3	底座相机0
base_1_camera_rgb_image	768×1024×3	底座相机1

其他特征

• timestamp: float32, shape [1]
• frame_index: int64, shape [1]
• episode_index: int64, shape [1]
• index: int64, shape [1]
• task_index: int64, shape [1]