New method generates 100 synthetic XR behavioral profiles from real data
arXiv paper shows how to expand XR datasets 100x at zero marginal cost.
Researchers Xiaozheng Wang and Ryan P. McMahan from the University of Texas at Dallas have published a novel approach to expand Extended Reality (XR) behavioral datasets without additional data collection. Their paper, submitted to arXiv on August 14, 2026, introduces a motion synthesis pipeline that combines Dynamic Time Warping (DTW) with trajectory interpolation to generate synthetic behavioral trajectories from existing XR motion data.
The team used the publicly available FAST VR assembly dataset to demonstrate their method, generating 100 synthetic behavioral trajectories and releasing them openly. When evaluated through motion-based user identification, hybrid datasets containing both real and synthesized trajectories achieved performance comparable to similarly sized real-only datasets. Crucially, the synthesized trajectories remained distinguishable from their contributing participants, preventing data leakage. This approach offers a scalable solution to the persistent challenge of limited XR behavioral data, potentially enabling larger-scale behavioral modeling and machine learning evaluation in immersive systems.
- Pipeline combines DTW and trajectory interpolation to generate synthetic XR motion data
- 100 synthetic trajectories released from FAST VR assembly dataset, expanding dataset size while preserving task structure
- Hybrid datasets with synthetic data matched real-data performance in user identification tasks
Why It Matters
Solves the XR data scarcity problem by generating synthetic behavioral populations at zero marginal cost.