潮汐特征识别在权值优化中的应用  

Application of tidal feature recognition in weight optimization

作　　者：任高强 杨恩源 侯文娟 朱朋 李钦 杨欢 REN Gao-qiang;YANG En-yuan;HOU Wen-juan;ZHU Peng;LI Qin;YANG Huan(China Mobile Group Design Institute Co.,Ltd.He'nan Branch,Zhengzhou 450000,China;China Mobile Group He'nan Co.,Ltd.,Zhengzhou 450000,China)

机构地区：[1]中国移动通信集团设计院有限公司河南分公司,郑州450000 [2]中国移动通信集团河南有限公司,郑州450000

出　　处：《电信工程技术与标准化》2024年第2期5-8,共4页Telecom Engineering Technics and Standardization

摘　　要：5G的天线权值技术应用比较成熟,现有的权值方案遵循常规的网参优化模型,归为静态权值。但是实际网络的各项指标是动态变化的,若能识别网络用户动态分布特征,将权值优化方案按照小时级切片,即可实现网随业动。如何识别网络用户动态分布特征,成为新的研究课题。本文提供了一种根据历史用户分布时空序列信息识别潮汐场景小区用户分布规律性变化的方案,为动态权值优化提供场景识别。The application of 5G antenna weight technology is relatively mature,but the existing weight schemes also follow conventional network parameter optimization models and is classified as static weights.However,the actual network is dynamically changing,by slicing the weight optimization scheme at an hourly level,it is possible to achieve network movement with industry movement.How to identify the dynamic distribution characteristics of network users has become a new research topic.This article provides a solution that can identify the regular changes in the distribution of users in tide scenario cells based on historical user distribution spatiotemporal sequence information,providing scene recognition for dynamic weight optimization.

关 键 词：潮汐权值方案 动态权值 卷积神经网络 规律挖掘 

分 类 号：TN929.5[电子电信—通信与信息系统]