LiHRA
收藏资源简介:
LiHRA是一个针对人机交互(HRI)场景设计的全新数据集,旨在促进自动化、基于学习的或经典的危险监测(RM)方法的发展。该数据集结合了3D激光雷达点云、人体关键点以及机器人关节状态,捕捉人机协作的完整空间和动态环境。数据集涵盖了六种代表性的人机交互场景,包括协作和共存任务、物体传递和表面抛光等,每个场景都包含安全和危险两种版本。LiHRA为训练和评估经典和基于AI的危险监测算法提供了丰富的资源。
LiHRA is a novel dataset designed for human-robot interaction (HRI) scenarios, aiming to advance the development of automated, learning-based, or classical risk monitoring (RM) methods. This dataset integrates 3D LiDAR point clouds, human body keypoints, and robot joint states to capture the complete spatial and dynamic environment of human-robot collaboration. The dataset covers six representative HRI scenarios, including collaborative and coexisting tasks, object handover, surface polishing and other typical tasks, with both safe and hazardous versions for each scenario. LiHRA provides rich resources for training and evaluating both classical and AI-based risk monitoring algorithms.
LiHRA: A LiDAR-Based HRI Dataset for Automated Risk Monitoring Methods
数据集概述
LiHRA是一个基于LiDAR的数据集,专为人机交互(HRI)中的自动化风险监测(RM)方法设计。该数据集提供3D LiDAR点云、人体关键点和机器人关节状态,捕捉真实世界HRI动态。数据集包含安全关键场景,如物体交接和共享工作空间交互,涵盖有意接触和无意碰撞,以支持风险监测和缓解研究。
数据集内容
- 6个场景:协作(危险、非危险)、物体交接(危险、非危险)、共存(危险、非危险)
- 4,431个点云:以10 Hz频率记录
- 机器人状态:机器人关节状态、机器人ROS 2 tf-Frames
- 人体关键点:作为ROS 2 tf-Frames提供,包含5个人体关键点
硬件信息
- LiDAR:Seyond Falcon Kinetic LiDAR
- 机器人:Franka Emika Robot (FER)
- 追踪器:HTC VIVE Tracker 3.0
使用许可
数据集仅用于研究目的,不得用于商业用途。
相关资源
- 论文:arXiv Dataset
- 网站许可证:Creative Commons Attribution-ShareAlike 4.0 International License




