移动边缘计算下的流媒体服务QoE计算方法研究

在移动边缘计算环境下，用户和边缘节点产生的海量动态数据之间存在交叉影响且异构数据难以融合计算，导致流媒体服务的数据处理和QoE计算的有效性和时效性严重不足。因此，针对移动边缘计算环境下流媒体服务QoE计算的准确性不高、效率较低、不能充分满足用户个性化需求等问题，本项目从用户端的个性化分析、服务端的质量自适应以及基于多端协同的流媒体服务QoE评价出发，结合移动边缘计算特性开展跨层数据感知和数据融合计算，重点研究基于多源异构数据的用户偏好计算方法、面向QoE的流媒体服务自适应机制以及基于混合网络的流媒体服务QoE评价方法。在此基础上，构建一种移动边缘计算环境下准确、高效、个性化满足的流媒体服务QoE综合计算方法，并搭建一个开放可视化的流媒体服务QoE综合计算平台。通过理论分析和实验平台验证所提方法及模型的性能和可靠性，为未来QoE保障的边缘服务生态构建提供理论依据、方法保障和技术支持。

In mobile edge computing environment, users and edge nodes generate massive dynamic data which has cross-effects and it is difficult to conduct heterogeneous data fusion, leading to low efficiency of data processing and QoE calculation for streaming services. Existing mobile edge computing based QoE calculation approaches are not accurate, efficient or personalized enough. Therefore, we elaborate the QoE research for streaming services from three aspects: user-level personalized analysis, service-level quality adaption, and multi-level collaboration based QoE evaluation, combined with the characteristic analysis and cross-layer data perception of mobile edge computing. By studying the key approaches, such as heterogeneous data fusion based user preference calculation, QoE-aware streaming service adaption, and hybrid network based QoE evaluation, the study proposes an accurate, efficient, and personalized QoE calculation method and a public visual platform for streaming services in mobile edge computing environment, and finally verify the reliability and performance of the above models and algorithms through theory analysis and simulation experiments. The proposal is expected to provide a solid theory foundation and technical support for the construction of QoE-aware edge service ecosystem.