Towards Network-friendly and Privacy-preserving Immersive Computing

Immersive computing has transformed the way we interact with digital and physical worlds by blurring boundaries between them. However, as these technologies evolve, they introduce growing challenges related to increasing network demands and heightened privacy concerns. This article addresses these issues through three pivotal studies. First, we conduct a comprehensive measurement study of commercial social virtual reality platforms and immersive telepresence systems. We then propose MagicStream, a novel system for reducing bandwidth consumption and latency in immersive telepresence leveraging semantic communication. Finally, we introduce MetaFL, a federated-learning-based framework to accurately authenticate extended reality (XR) users while preserving their privacy. Our extensive evaluations reveal their advances in network efficiency and privacy compared to state-of-the-art efforts.