randi233/05iro-plus

该数据集是一个用于机器人学领域的数据集，专门针对so_follower机器人类型，通过LeRobot项目创建。它包含156个episodes和140325帧数据，用于训练和评估机器人控制任务。数据集结构包括动作特征（如关节位置：shoulder_pan.pos、shoulder_lift.pos等）、观测状态（与动作相同的关节位置）、前视图像（分辨率为480x640的RGB视频，每秒30帧），以及时间戳、帧索引、episode索引等元数据。数据以Parquet格式存储，视频以MP4格式存储，适用于机器人模仿学习或强化学习应用。This dataset is designed for robotics applications, specifically for the so_follower robot type, created using the LeRobot project. It contains 156 episodes and 140325 frames, intended for training and evaluating robot control tasks. The dataset structure includes action features (e.g., joint positions: shoulder_pan.pos, shoulder_lift.pos, etc.), observation states (same joint positions as actions), front-view images (RGB videos with 480x640 resolution at 30 fps), and metadata such as timestamps, frame indices, episode indices, etc. Data is stored in Parquet format, with videos in MP4 format, suitable for robot imitation learning or reinforcement learning applications.