Research And Implementation Of The Mobile Social Virtual Reality

Benefiting from Virtual Reality's(VR)revolutionary immersive user experience,social VR is not restricted to instant messaging,video chat and so on,but let users have face-to-face social activities in a shared virtual environment.To pursue better mobility and lower prices,both big companies and startups launch their own mobile social VR applications on smartphone to enter this brand new market.However,the research community still knows little about how and how well mobile social VR works.And the lack of such understanding makes it hard for people to propose further optimizations and better deployments.To lay a good foundation for better design choices of mobile social VR,we need to have a thorough understanding of both mobile VR and social VR.Thus,this paper first presents a comprehensive analysis and measurement study on today's mobile VR systems.I employ mobile VR system's SDK to discover how smartphone provides immersive VR experience and summarize it as mobile VR applications' technical framework.Meanwhile,this paper also profiles popular mobile VR applications to get a picture of how well can they offer the user experience.The experimental results reveal that most existing mobile VR applications cannot provide satisfying user experience,and smartphone's constrained CPU/GPU capability and weak battery are the bottlenecks.Based upon the analysis and experiments,this paper discusses main challenges and potential prospects for mobile VR.Further aiming at mobile social VR,this paper identifies three independent features from the complicated workflow provided by commercial social VR applications.Taking the three features as design objective,this paper propose MSVR,a system design for mobile social VR.To offer good compatibility and easy deployment,MSVR utilizes existing well-known network techniques to accomplish essential communications in social VR.Besides,MSVR can also hide significant network delay to avoid degrading the whole system's responsiveness.In order to validate MSVR's functionality and analyze its performance,this paper implements MSVR as a prototype and evaluates both system performance and resource usage.Specifically,this paper decouples the MSVR prototype to evaluate social VR's features individually.Through the feature characterization,I obtain many social VR-specific conclusions.Such as social VR's performance is mainly dependent on the complexity of the virtual scene it shares.Those social VR-specific findings are valuable to future work in this area.