Xense/shoe_insole_retrieval_and_packing0515

该数据集是使用LeRobot平台创建的双Flexiv Rizon4机器人操作数据集，专为机器人学习任务设计。数据集包含101个完整的情节（episodes），总计206,299帧，以30帧每秒的速率采集。数据以parquet文件和视频文件形式存储，总数据文件大小为100MB，视频文件大小为500MB。数据集仅包含训练集，涵盖一个任务类型。特征包括：动作数据（20维浮点数组，表示左右机器人的TCP位姿和夹爪位置）、状态观测（与动作相同的20维结构）、多视角图像观测（包括头部摄像头、左右手腕摄像头以及左右触觉传感器图像，分辨率分别为480x640和400x700，均为RGB视频），以及时间戳、帧索引、情节索引等元数据。该数据集适用于机器人控制、模仿学习或强化学习研究。This dataset is a dual Flexiv Rizon4 robot manipulation dataset created using the LeRobot platform, designed for robotics learning tasks. It contains 101 complete episodes with a total of 206,299 frames, captured at 30 frames per second. The data is stored in parquet and video files, with a total data file size of 100MB and video file size of 500MB. The dataset includes only a training split and covers one task type. Features include: action data (20-dimensional float array representing TCP poses and gripper positions for both left and right robots), state observations (same 20-dimensional structure as actions), multi-view image observations (including head camera, left/right wrist cameras, and left/right tactile sensor images, with resolutions of 480x640 and 400x700 respectively, all as RGB videos), and metadata such as timestamps, frame indices, and episode indices. This dataset is suitable for robot control, imitation learning, or reinforcement learning research.