边缘计算驱动的QoE可感知360度全景流媒体分发机制研究

随着5G时代的到来，以360度全景视频为核心支撑的VR/AR类应用将成为互联网的下一个热点。然而，由于360度全景视频内在的高带宽、低时延的需求，在现有互联网上大规模传输分发360度全景视频面临巨大的挑战。本项目拟研究设计一种用户体验（QoE）可感知的自适应全景流媒体分发机制，提高面向头戴式终端的全景视频流媒体分发质量，优化用户的全景流媒体感知体验，更好地支持超高清全景流媒体内容的分发。为突破高服务质量保障的全景流媒体传输的难点问题，本项目将提出一种边缘计算驱动的新型全景流媒体分发网络体系结构，建立用户QoE评价和保障理论框架，设计全景视频传输关键算法，优化任务迁移与内容分发耦合机制，搭建用户数据采集平台，并开发端到端全景流媒体原型系统。本项目的研究成果将有助于加快VR/AR相关产业场景的应用落地（如教育、医疗、娱乐等），推动我国信息产业的发展。

In the coming 5G era, VR/AR applications built on 360-degree video streaming will become the next blue ocean of the Internet. However, due to the inherent requirements on high bandwidth and low latency, it is inevitably a great challenge to design and deploy the large-scale transmission and distribution of 360-degree videos on the existing Internet structure. This project plans to research and design a user quality-of-experience (QoE) adaptive 360-degree video streaming distribution mechanism. The ultimate goal is to improve the QoE of users watching 360-degree video streaming media, and support the distribution of ultra high definition (UHD) 360-degree video streaming distribution. Facing the demanding low latency for transmitting the UHD contents in limited bandwidth/computing resources, this project will propose a new 360-degree streaming system architecture by utilizing edge computing. The key research issues include establishing the theoretical framework of user QoE evaluation, designing an efficient 360-degree video streaming transmission strategy, optimizing the coupling mechanism design of task migration and content distribution, and finally developing an end-to-end 360-degree video streaming system prototype. The research results of this project will help to accelerate the application of VR/AR related industrial scenarios (such as education, health care, entertainment, etc.), and promote the development of China's information industry.