dual_arm_ur5e_ur3e_20260530_v3

该数据集是一个用于机器人学的双臂操作数据集，基于LeRobot平台创建。它包含两个机械臂（UR5e和UR3e）的交互数据，总共有2个episode、1094帧和1个任务。数据以15fps的帧率采集，存储为Parquet格式。观测特征包括三个视频流：base_image、ur5e_image和ur3e_image，每个视频的分辨率为224x224x3，编码为h264，无音频。状态观测（observation.state）是一个14维浮点向量，表示两个机械臂的关节位置和夹爪状态（UR5e的6个关节和1个夹爪，UR3e的6个关节和1个夹爪）。动作（action）也是一个14维浮点向量，对应两个机械臂的动作控制。此外，数据集还包含时间戳、帧索引、episode索引、索引和任务索引等元数据。数据集结构支持训练分割（train: 0:2），数据文件和视频文件总大小均约为0.0001MB。This dataset is a dual-arm manipulation dataset for robotics, created using the LeRobot platform. It contains interaction data from two robotic arms (UR5e and UR3e), with a total of 2 episodes, 1094 frames, and 1 task. The data is collected at a frame rate of 15fps and stored in Parquet format. Observation features include three video streams: base_image, ur5e_image, and ur3e_image, each with a resolution of 224x224x3, encoded in h264, and without audio. The state observation (observation.state) is a 14-dimensional floating-point vector representing the joint positions and gripper states of the two robotic arms (6 joints and 1 gripper for UR5e, and 6 joints and 1 gripper for UR3e). The action is also a 14-dimensional floating-point vector corresponding to the motion control of the two robotic arms. Additionally, the dataset includes metadata such as timestamps, frame indices, episode indices, indices, and task indices. The dataset structure supports a training split (train: 0:2), with the total size of data files and video files both approximately 0.0001MB.