cube_swap_term_project_v4

该数据集是一个基于LeRobot创建的机器人数据集，专门针对双Piper跟随器机器人（bi_piper_follower）。它包含20个episodes，总计23777帧，覆盖1个任务。数据以parquet文件格式存储动作和状态信息，视频文件为mp4格式，帧率为30fps。数据集的特征包括：动作数据（一个14维浮点数组，表示左右机械臂的6个关节位置和夹爪位置）、观察状态（同样是一个14维浮点数组，反映机器人关节和夹爪的实时位置）、来自多个摄像头的图像观察（包括左主摄像头（分辨率720x1280）、左腕部摄像头（分辨率480x640）和右腕部摄像头（分辨率480x640），均为RGB视频），以及时间戳、帧索引、episode索引等元数据。该数据集适用于机器人控制、模仿学习或强化学习任务，旨在提供机器人操作的多模态数据（如动作、状态和视觉信息）。

This dataset is a robotics dataset created using LeRobot, specifically designed for the bi_piper_follower robot. It contains 20 episodes, totaling 23777 frames, covering 1 task. The data is stored in parquet file format for action and state information, with video files in mp4 format at a frame rate of 30fps. The dataset features include: action data (a 14-dimensional floating-point array representing the joint positions of the left and right robotic arms and gripper positions), observation state (also a 14-dimensional floating-point array reflecting the real-time positions of robot joints and grippers), image observations from multiple cameras (including a left main camera with resolution 720x1280, a left wrist camera with resolution 480x640, and a right wrist camera with resolution 480x640, all in RGB video format), as well as metadata such as timestamps, frame indices, and episode indices. This dataset is suitable for robotics control, imitation learning, or reinforcement learning tasks, aiming to provide multimodal data for robot operations (e.g., actions, states, and visual information).