xhaka3456/aic_ur5e_sim_v4_tc_v1cmd_756eps

这是一个用于机器人学习的数据集，由LeRobot创建。数据集使用UR5e机器人收集，包含756个episodes和204,272帧，涉及3个不同的任务。数据以parquet格式存储，视频以mp4格式存储，帧率为20fps。数据集的特征包括来自左、中、右三个摄像头的图像观测（分辨率480x640，RGB三通道），机器人状态（26维，包括TCP位置、姿态、速度、关节角度、力和扭矩等），动作（6维，表示delta移动和轴角旋转），以及时间戳、帧索引、episode索引等元数据。数据集主要用于训练和评估机器人控制算法，支持模仿学习、强化学习等任务。This is a dataset for robotics learning, created using LeRobot. It was collected with a UR5e robot, containing 756 episodes and 204,272 frames across 3 different tasks. The data is stored in parquet format, with videos in mp4 format at 20fps. The dataset features include image observations from left, center, and right cameras (resolution 480x640, RGB three channels), robot state (26-dimensional, covering TCP position, orientation, velocity, joint angles, force, and torque), actions (6-dimensional, representing delta movements and axis-angle rotations), and metadata such as timestamp, frame index, episode index. It is primarily used for training and evaluating robot control algorithms, supporting tasks like imitation learning and reinforcement learning.