TEXR: The University of Texas Extended Reality Headset Tracking Dataset

The next generation of XR devices will require seamless functionality in any environment and require globally referenced poses. This paper introduces a public benchmark dataset for extended reality (XR) headset tracking in a variety of outdoor environments. The dataset includes raw GNSS intermediate frequency samples, timestamped measurements from three inertial measurement units, and globally shuttered stereoscopic cameras, all collected on a head-mounted sensor platform. Our benchmark dataset recordings include data collected in urban, light-urban, and open field environments. The dataset includes differing levels of user dynamics corresponding to pedestrian and gaming use cases. Similar datasets exist for ground and aerial vehicle-mounted sensor platforms. However, no dataset currently exists for head-mounted XR platforms with this combination of sensors. This dataset enables the evaluation of computer vision and sensor fusion algorithms for headset tracking on real-world data.