TH ¨OR

TH ¨OR数据集是由瑞典厄勒布鲁大学的移动机器人与嗅觉实验室创建的，专注于室内环境中的人类移动轨迹和眼动数据。该数据集包含超过60分钟的高频（100 Hz）记录，总计395k帧，捕捉了2531k次人类检测和600多个个体及群体的移动轨迹。数据收集过程中，参与者在模拟的社交环境中与机器人互动，执行各种任务，如导航、物品搬运等。TH ¨OR数据集不仅记录了位置和方向，还包括了障碍物地图、目标坐标、社交群体信息等丰富内容。此数据集旨在支持人机交互、移动机器人导航和人类行为预测等领域的研究，提供了一个高质量、多维度的数据资源，以促进相关技术的开发和测试。

The THÖR Dataset was created by the Mobile Robotics and Olfaction Laboratory at Örebro University, Sweden, and focuses on human movement trajectories and eye-tracking data in indoor environments. This dataset contains over 60 minutes of high-frequency (100 Hz) recordings, totaling 395,000 frames, capturing 2,531,000 human detections and movement trajectories of over 600 individuals and groups. During data collection, participants interacted with robots in simulated social environments and performed various tasks such as navigation and object transportation. In addition to recording position and orientation, the dataset also includes rich supplementary content such as obstacle maps, target coordinates, and social group information. This dataset aims to support research in fields including human-robot interaction, mobile robot navigation, and human behavior prediction, providing a high-quality, multi-dimensional data resource to promote the development and testing of relevant technologies.