Yeshun-weston/so101-BlueCube-BlackMat-50epi

该数据集是一个机器人控制数据集，专门用于机器人跟随任务（robot_type: so_follower）。它包含50个episode，总计25954帧数据，专注于一个任务。数据集结构包括动作（action）和观测（observation）特征：动作特征为6维浮点数组，表示机器人关节位置（如肩部平移、肩部提升、肘部弯曲、手腕弯曲、手腕旋转和夹爪位置）；观测特征包括状态观测（同样为6维关节位置）和前端摄像头图像观测（视频格式，分辨率480x640，30帧/秒，彩色图像）。此外，还包含时间戳、帧索引、episode索引等元数据。数据集使用LeRobot工具创建，适用于机器人学习和控制研究，特别是模仿学习或强化学习场景。

This dataset is a robotics control dataset designed for robot following tasks (robot_type: so_follower). It consists of 50 episodes with a total of 25,954 frames, focusing on a single task. The dataset structure includes action and observation features: the action feature is a 6-dimensional float array representing robot joint positions (e.g., shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper positions); observation features include state observation (also 6-dimensional joint positions) and front camera image observation (video format, resolution 480x640, 30 fps, color images). Additionally, it contains metadata such as timestamp, frame index, episode index, etc. The dataset was created using the LeRobot tool and is suitable for robotics learning and control research, particularly in imitation learning or reinforcement learning scenarios.