摘要
中国移动主导的VoLTE视频业务作为当前拉动收入的重要增长点,对带宽、时延、分组丢失均有较高的要求,其基于应用和感知的特性导致业务整体评估和优化难度较大.且目前业界缺乏VoLTE视频感知评估的有效方案,已有的ITU-T的G.1070和U-VMOS等视频感知评估方法基本无法适用VoLTE视频感知评估.提出了一种基于BP神经网络的VoLTE视频质量评估方法,可以客观评价实时VoLTE视频质量.同时提出一种VoLTE视频“双向四维三阶”端到端问题定界方法,准确定位VoLTE视频质量原因,可以有效提升VoLTE视频用户的感知评估准确率和优化效率.
VoLTE is a service of the application layer,which has become a new growth engine of China Mobile.How to measure customers’perception in VoLTE is difficult,while currently relevant standards(ITU-T G.1 070 and U-VMOS)cannot meet this new challenge.Network Optimizing also need to be increasingly complex and satisfy the requirements of a high-performance system,for example,wider bandwidth,less latency and lower loss.This paper focus on a network evaluation scheme and a partition algorithm for VoLTE quality problem,increasing the effectiveness and efficiency of both evaluation and improving.The evaluation scheme based on BP neural network can grade VoLTE real-time and provide feedback to network optimization.The partition algorithm can lead to identification,analysis and troubleshooting of performance issues,which is implemented on the platform and achieve very good results.
作者
张晓
张弘
孙斌
徐小龙
ZHANG Xiao;ZHANG Hong;SUN Bin;XU Xiao-long(China Mobile Group Gansu Co.,Ltd.,Lanzhou 730070,China)
出处
《电信工程技术与标准化》
2018年第A01期29-32,共4页
Telecom Engineering Technics and Standardization
