← All Projects 
AR/XR · Tracking
Device-Agnostic Multi-Sensor Tracking for AR
Published
A device-agnostic framework for surgical tool tracking in AR that combines HoloLens 2 and optical tracker observations through a dynamic scene graph and pose-graph optimization to reduce tracking error and loss rate across heterogeneous sensors.
Contributions
- Developed the dynamic scene graph for multi-sensor tracking state management.
- Implemented pose-graph optimization for fusing HoloLens 2 and optical tracker observations.
- Evaluated tracking accuracy and loss rate in surgical AR scenarios.