cell2_20260520_hossam_coffee-shop_setting_cycle_20260520_204102

该数据集是一个机器人操作数据集，使用LeRobot平台构建，专门针对starpilot_yam_gripper双机械臂系统。数据采集于咖啡店场景（cell2_20260520_hossam_coffee-shop_setting），包含8个完整episodes，总计3402帧，覆盖一个任务。数据集提供丰富的多模态信息：动作空间包含14维浮点数，用于控制双机械臂的末端执行器位置、旋转和夹爪宽度；状态观测包括32维浮点数，涵盖机械臂的编码器角度、IMU线性加速度与角速度、末端执行器位姿（位置和四元数）、夹爪开合状态和距离；视觉观测包括来自左腕、右腕和基座摄像头的RGB图像（部分带有深度信息），分辨率分别为480x640和768x1024，帧率为30fps。数据以parquet文件（100MB）和视频文件（200MB）形式存储，适用于机器人学习、模仿学习或强化学习任务。This dataset is a robot manipulation dataset created using the LeRobot platform, specifically designed for the starpilot_yam_gripper dual-arm system. Data was collected in a coffee shop scenario (cell2_20260520_hossam_coffee-shop_setting), containing 8 complete episodes with a total of 3402 frames, covering one task. The dataset provides rich multimodal information: the action space consists of 14-dimensional floating-point numbers controlling the end-effector position, rotation, and gripper width of the dual arms; state observations include 32-dimensional floating-point numbers covering the arm encoder angles, IMU linear acceleration and angular velocity, end-effector pose (position and quaternion), gripper open/close state, and distance; visual observations include RGB images from left wrist, right wrist, and base cameras (some with depth information), with resolutions of 480x640 and 768x1024, at a frame rate of 30fps. Data is stored in parquet files (100MB) and video files (200MB), suitable for robot learning, imitation learning, or reinforcement learning tasks.