RoboCOIN/AI2_Alphabot_2_place_pen
收藏资源简介:
该数据集名为AI2_Alphabot_2_place_pen,是一个机器人操作数据集,专注于将笔放入笔筒中的任务。它基于LeRobot格式并完全兼容LeRobot框架。数据集包含527个episodes,总计666,384帧,大小为13.18 GB,帧率为30 FPS。使用AI2_Alphabot_2机器人,配备双指末端执行器,并通过VR控制器进行遥操作。传感器包括四个RGB摄像头:前胸摄像头、前头摄像头、左腕摄像头和右腕摄像头,分辨率均为640x480。场景设置为教育->学校环境,涉及物体包括白色笔筒、黑色圆珠笔和黑色钢笔。数据包含状态信息(34维,如关节角度、末端执行器位姿)和动作信息(34维),并提供丰富的注释,如末端执行器加速度、方向、速度等。数据集由北京智源人工智能研究院的RoboCOIN团队贡献,用于机器人学习和操作研究。
This dataset, named AI2_Alphabot_2_place_pen, is a robotics manipulation dataset focused on the task of placing the pen into the pen holder. It uses an extended format based on LeRobot and is fully compatible with LeRobot. The dataset contains 527 episodes, totaling 666,384 frames, with a size of 13.18 GB and a frame rate of 30 FPS. It employs the AI2_Alphabot_2 robot with a two-finger end-effector and VR controller teleoperation. Sensors include four RGB cameras: cam_front_chest_rgb, cam_front_head_rgb, cam_left_wrist_rgb, and cam_right_wrist_rgb, all with a resolution of 640x480. The scene is set in an Education->School environment, with objects including a white pen holder, a black ballpoint pen, and a black fountain pen. The data includes state information (34-dimensional, such as joint angles and end-effector poses) and action information (34-dimensional), along with rich annotations like end-effector acceleration magnitude, direction, and velocity. Contributed by the RoboCOIN Team at Beijing Academy of Artificial Intelligence (BAAI), it is designed for robot learning and manipulation research.




