HuskyMango/5-28
收藏资源简介:
该数据集是一个机器人技术数据集,专为单臂Franka机器人设计,用于机器人学习和控制任务。数据集包含1个完整任务(episode),总计243帧数据,以20帧每秒(fps)的速率采集。数据以parquet格式存储,总数据文件大小为100MB,视频文件大小为200MB。数据集结构包括动作特征(action),如机器人末端执行器的位置(x, y, z)、四元数姿态(qx, qy, qz, qw)和夹爪状态(gripper),以及观察状态特征(observation.state),涵盖机器人7个关节的位置、夹爪状态和深度信息(从depth_0000到depth_0187的x、y、z坐标),用于提供环境感知数据。数据集适用于训练和评估机器人控制算法,支持强化学习或模仿学习应用。
This dataset is a robotics dataset designed for a single-arm Franka robot, intended for robot learning and control tasks. It contains 1 complete task (episode) with a total of 243 frames, captured at a rate of 20 frames per second (fps). The data is stored in parquet format, with a total data file size of 100MB and video file size of 200MB. The dataset structure includes action features, such as robot end-effector position (x, y, z), quaternion orientation (qx, qy, qz, qw), and gripper state, as well as observation state features, covering the positions of 7 robot joints, gripper state, and depth information (x, y, z coordinates from depth_0000 to depth_0187) for environmental perception. The dataset is suitable for training and evaluating robot control algorithms, supporting applications in reinforcement learning or imitation learning.




