Scalable Collocated Multi-User VR Through Virtual Environment User Spatial Coherence

Area of Science:

• Computer Science
• Virtual Reality

Background:

• Server-client communication bottlenecks hinder scalability in collocated multi-user virtual reality (VR).
• Existing unicast approaches struggle to efficiently support large groups due to individual data streams.

Purpose of the Study:

• To present an efficient multicast-based communication strategy for collocated multi-user VR.
• To improve the speed and completeness of environment reconstruction in large VR groups.

Main Methods:

• Implemented a system that multicasts identical packets to all clients, employing packetization, visibility-aware prioritization, and lightweight acknowledgments.
• Partitioned the environment into independently decodable packets, with transmission order guided by precomputed visibility footprints.
• Utilized repeated multicast and compact bitmap acknowledgments for scalability, ensuring communication cost is largely independent of client numbers.

Main Results:

• The multicast approach demonstrated significantly faster environment reconstruction compared to unicast TCP-based methods.
• Fewer missing environment parts were observed with the multicast system in controlled studies.
• The system successfully supported collocated VR groups of 7 and 16 participants.

Conclusions:

• Multicast transmission over commodity Wi-Fi is a viable solution for large collocated VR groups.
• The proposed method effectively alleviates server-client communication bottlenecks in multi-user VR environments.
• This approach enhances efficiency and fairness in VR data delivery.