RoboCOIN/AI2_Alphabot_2_package_fruit_1
收藏资源简介:
该数据集是一个名为AI2_Alphabot_2_package_fruit_1的机器人数据集,使用基于LeRobot的扩展格式,并完全兼容LeRobot框架。它专注于一个具体的机器人操作任务:从桌子上拾取草莓和芒果,并将它们放入一个棕色篮子中。数据集包含309个完整的情节(episodes),总计138,843帧视频数据,帧率为30 FPS,总大小为2.15 GB。使用的机器人是AI2_Alphabot_2,配备双指末端执行器,通过VR控制器进行遥操作。场景设置在零售超市环境中。数据集提供了多视角RGB视频观测,包括胸前前置、头部前置、左手腕和右手腕四个摄像头,每个视频分辨率为640x480,使用h264编解码器。此外,数据集还包含机器人的状态数据(如关节位置、末端执行器位姿)和动作数据,维度均为34维,以及丰富的时间戳、帧索引和注释信息。数据以Parquet文件格式组织,并包含训练集分割。该数据集由北京人工智能研究院(BAAI)的RoboCOIN团队贡献,旨在支持机器人操作和模仿学习的研究。
This dataset, named AI2_Alphabot_2_package_fruit_1, uses an extended format based on LeRobot and is fully compatible with LeRobot. It focuses on a specific robotic manipulation task: picking a strawberry and a mango from a table and placing them into a brown basket. The dataset comprises 309 total episodes, with 138,843 total frames at 30 FPS, and has a total size of 2.15 GB. The robot used is the AI2_Alphabot_2, equipped with a two-finger end-effector and teleoperated via a VR controller. The scene is set in a retail supermarket environment. It provides multi-view RGB video observations from four cameras: front chest, front head, left wrist, and right wrist, each with a resolution of 640x480 and using the h264 codec. Additionally, the dataset includes robot state data (e.g., joint positions, end-effector poses) and action data, both with 34 dimensions, along with rich timestamp, frame index, and annotation information. The data is organized in Parquet files and includes a training split. This dataset is contributed by the RoboCOIN Team at the Beijing Academy of Artificial Intelligence (BAAI) and is intended to support research in robotic manipulation and imitation learning.




