Modeling the tele-immersive systems using stochastic activity network

The next-generation social communication medium, tele-immersion, is receiving increasing attention from both the research and industrial communities. It enables interaction between geographically distributed sites through realistic reconstruction of physical scenes in immersive spaces. While the previous research focused on real-world measurements of implemented tele-immersive systems, we take a model-based approach to study the system more generally and thoroughly. With a model validated with actual measurements, we evaluate the system performance in a larger scale, and understand the quantitative effect of the algorithmic enhancements that were proposed in the literature. We also predict futuristic performance characteristic and show that high interactivity can be foreseen with hardware advances in the near future.