g1_brack_pickbottle_dataset

该数据集是一个机器人学数据集，使用LeRobot工具创建，专门针对Unitree_G1_Brainco双臂机器人。它包含201个训练集，总计174,483帧和804个视频，帧率为30fps。数据集提供了丰富的特征，包括机器人的26维状态观察（如左右肩、肘、腕和手部关节角度）和26维动作控制信号，以及来自四个摄像头的图像数据：左高摄像头、右高摄像头、左腕摄像头和右腕摄像头，每个摄像头输出分辨率为480x640的RGB视频。此外，还包含时间戳、帧索引、集索引等元数据，适用于机器人控制、模仿学习或强化学习任务。This dataset is a robotics dataset created using the LeRobot tool, specifically designed for the Unitree_G1_Brainco dual-arm robot. It contains 201 training sets, totaling 174,483 frames and 804 videos, with a frame rate of 30fps. The dataset provides rich features, including the robots 26-dimensional state observations (such as left and right shoulder, elbow, wrist, and hand joint angles) and 26-dimensional action control signals, as well as image data from four cameras: left high camera, right high camera, left wrist camera, and right wrist camera, each outputting RGB videos with a resolution of 480x640. Additionally, it includes metadata such as timestamps, frame indices, and episode indices, making it suitable for robotics control, imitation learning, or reinforcement learning tasks.