用户参与度影响下的实时视频QoE评估及跨层传输方法研究

Video services are of great significance to national security, national economy and people's lives. With the emergence of a large number of the rising video services, the users are no longer the simple video receivers but can participate in all aspects of the video playback. Due to this, the subjective initiative of the users is enlarged. It has great values both on the theory and practice to explore the relationship between the user participation based QoE assessment and the improvement of the satisfaction of the video users. To this end, we focus on the research of the user participation based real-time video QoE assessment and cross layer optimization, and plan to reveal the relationship between the user participation, the video content, the wireless network state, the playback information and the improvement of the user experience, as well as the regularity of the cross layer parameter optimization. Then the user participation based video QoE model is developed, and a QoE driven cross layer optimization for video transmission is derived and solved. To validate the effectiveness of the proposed model and the optimization method, we build a simulation platform for real video to realize the real-time video QoE assessment and cross layer transmission. Innovation points: The users' personalized features and the new video services' characteristics are introduced into the video quality assessment system with the user at the core and the development of the video services. We formulate a global QoE driven cross layer optimization problem for video transmission, which involves the user layer, the application layer and the layers of the wireless network, to optimize the parameters' configuration of all the layers simultaneously. The proposed video QoE model and optimization method are validated and modified by our built simulation platform for real video.

视频应用对国家安全、国民经济和人民生活意义重大。随着大量新兴视频业务的涌现，用户不再是单纯的视频信息接收者，而是能够参与到视频播放的各个环节，主观能动性不断放大。探究用户参与度与视频体验质量(QoE)之间的作用机理对提升视频用户满意度具有理论意义和应用价值。本项目以用户参与度影响下的实时视频QoE评估和跨层传输为研究对象，阐明用户参与度、视频内容、网络特征及播放状态等对用户体验的影响机理，揭示跨层参数配置与视频QoE提升之间的内在规律，建立基于用户参与度的视频QoE模型，构造并求解QoE驱动的视频跨层传输优化问题，开发真实视频流QoE评估和跨层传输仿真系统。创新点：以用户为核心、视频业务发展为导向，将用户特性和新兴视频业务特性有机融合，建立新的视频QoE评估体系；构造视频QoE驱动的全局统筹跨层优化问题，同时对多层参数进行优化；搭建真实视频流仿真系统对所提模型和算法进行性能验证和参数修正。

随着用户智能终端设备的普及和新一代无线通信技术的蓬勃发展，超高体验的新兴视频应用大量涌现，如何合理有效的对新兴的视频应用体验进行合理有效的评估以及在新一代通信系统中提升新兴视频应用的体验质量具有重要意义。本项目为用户参与度影响下的实时视频QoE评估及跨层传输方法研究。主要研究内容分为以下三个部分：1、视频QoE驱动的跨层传输优化研究。在该部分内容中，从两个方面进行详细研究：播放终端状态感知的跨层优化和视频内容和播放状态联合感知的跨层优化。在播放终端感知的跨层优化方面，由于播放流畅性是影响视频QoE的关键因素，建立了播放状态与无线资源分配之间的映射模型，提出了DASH用户播放状态感知的跨层优化方法；在视频内容和播放状态联合感知的跨层优化方面，从视频压缩域出发，提出了VoD视频用户QoE驱动的联合视频编码参数和视频包紧急性的跨层优化方法；2、用户特性感知的视频QoE模型建立及应用研究。视频用户QoE模型是本项目研究的关键，为了准确评估不同应用场景下视频用户QoE模型，在该部分内容中，从两个方面进行研究：LTE网络下DASH视频用户QoE模型和云边协同网络下基于移动边缘计算技术的实时直播/点播用户QoE模型建立及应用。前者从视频帧角度出发，构造了综合考虑视频接收质量、视频帧质量波动、终端播放流畅性的DASH用户QoE模型并将其应用于DASH视频传输跨层优化。后者综合考虑用户本身的属性以及用户在视频播放阶段的参与度，建立了用户属性、参与度、视频内容以及无线网络参数联合感知的视频QoE模型，并将其应用于云边协同网络中的移动边缘计算缓存跨层优化。3、真实视频流QoE评估和跨层优化仿真系统的建立。已搭建真实视频流仿真系统，实现了对所提模型和算法进行性能验证和参数修正。