4G/5G语音业务质差小区判别关键技术研究  

Research on Key Technologies for Identifying Poor-Quality Voice Service Cells in 4G/5G Network

作　　者：王勇[1] 赵振桥 刘喜卿[2] 程新洲[2] Wang Yong;Zhao Zhenqiao;Liu Xiqing;Cheng Xinzhou(China United Network Communications Group Co.,Ltd.,Beijing 100033,China;China Unicom Research Institute,Beijing 100048,China)

机构地区：[1]中国联合网络通信集团有限公司,北京100033 [2]中国联通研究院,北京100048

出　　处：《邮电设计技术》2025年第3期71-74,共4页Designing Techniques of Posts and Telecommunications

摘　　要：针对当前日趋复杂的网络架构及日益提升的网络优化难度,提出并实现一种新型的基于多域数据融合的算法系统。该系统通过关联IMS、5GC及EPC域的XDR数据,实现对VoLTE、VoNR及EPS FB等语音业务的精准区分。在此基础上,系统通过构建各业务的质量劣化阈值模型,并采用滑动窗口算法,从而精准筛选重点质差小区。经X市现网连续5天的实际数据验证表明,该系统能够有效识别重点质差小区,充分验证系统的可行性与实用性。In response to the growing complexity of network architectures and the challenges in network optimization,it proposes a novel algorithm system based on multi-domain data fusion.The system leverages the correlation of XDR data across IMS,5GC,and EPC domains to accurately distinguish VoLTE,VoNR and EPS FB voice services.Furthermore,it constructs servicespecific quality degradation threshold models and employs a sliding window algorithm to screen key under-performing cells with high precision.Through the data verification of a 5-day period in X city,the system can pinpoint key under-performing cells,thereby confirming its feasibility and practical applicability.

关 键 词：VoNR VoLTE EPS FB 网络智能运营 

分 类 号：TN915[电子电信—通信与信息系统]