CTR232

该数据集使用LeRobot创建，是一个专注于CTR（连续体机器人）操作的机器人学数据集。它包含232个episodes，总计98761帧，覆盖4个任务。数据以30fps采集，特征包括动作（36维，如电机控制、目标位置、触摸反馈等）、观测状态（58维，如正向运动学、电机反馈、触摸和NDI传感器数据）以及来自多个全局摄像头的图像观测（Cam1-Cam4为960x960分辨率，Cam5为400x400分辨率，均为RGB视频）。数据格式为Parquet文件，视频使用AV1编码，总数据大小约300MB（数据100MB，视频200MB）。该数据集适用于机器人控制、强化学习或计算机视觉任务的研究。This dataset is created using LeRobot and focuses on CTR (Continuum Robot) manipulation in robotics. It contains 232 episodes, totaling 98,761 frames, covering 4 tasks. Data is collected at 30fps, with features including actions (36-dimensional, such as motor control, target positions, and tactile feedback), observation states (58-dimensional, such as forward kinematics, motor feedback, tactile and NDI sensor data), and image observations from multiple global cameras (Cam1-Cam4 at 960x960 resolution, Cam5 at 400x400 resolution, all in RGB video). The data format is Parquet files, with videos encoded in AV1, and the total data size is approximately 300MB (100MB for data and 200MB for videos). The dataset is suitable for research in robot control, reinforcement learning, or computer vision tasks.