Reasearch Assistant
Research, Duke University I3T Lab, 2023.11 - 2024.07
- Skill Sets: C++/Python/Linux/SLAM/ROS/Swift/C#/Unity
- Designed and implemented a modular testbed that integrates a Vicon motion capture system with external camera support to evaluate the spatial tracking accuracy of XR headsets, such as Apple Vision Pro and Meta Quest 3, using open-source VI-SLAM algorithms as benchmarks.
- Developed robust Meta Quest 3 and Apple Vision Pro applications to enable synchronized sensor data collection and real-time tracking analysis. The system facilitated detailed performance evaluation under varied environmental conditions and user movements.
- Led comprehensive experiments leveraging the Vicon system to assess the impact of movement patterns, environmental features, and lighting conditions on XR tracking performance. These experiments revealed a 33.9% reduction in relative pose error and a 14.6% reduction in absolute pose error for the Apple Vision Pro compared to Meta Quest 3