一种基于GBDT机器学习的算法及应用研究  

Algorithm and application based on GBDT machine learning

作　　者：周徐 方东旭 文冰松 ZHOU Xu;FANG Dong-xu;WEN Bing-song(China Mobile Group Chongqing Co.,Ltd.,Chongqing 400021,China)

机构地区：[1]中国移动通信集团重庆有限公司

出　　处：《电信工程技术与标准化》2019年第11期27-30,46,共5页Telecom Engineering Technics and Standardization

摘　　要：本文采用GBDT机器学习的MassiveMIMO波束自适应配置算法,基于基站参数、地形地物、业务/用户分布等特征使用AI算法自适应匹配最佳的波束配置,同时建立实时定位跟踪引擎。通过5G终端MR及采样数据流结合MassiveMIMO波束自适应配置算法实时对终端进行跟踪定位,并根据无线环境变化进行波束优化调整,解决了MassiveMIMO海量模式选择性难题、实时优化、支撑工业互联网的5G实时优化平台开发等关键问题。This method adopts Massive MIMO beam adaptive configuration algorithm for GBDT machine learning,adaptive matching of optimal beam configuration with AI algorithm based on base station parameters、terrain feature、business/user profile and so on.At the same time,a real-time positioning and tracking engine is established,real-time tracking and positioning of a terminal through a 5 G terminal MR and a sampling data stream in combination with a passive MIMO beam adaptive configuration algorithm,according to the change of wireless environment,the beam is optimized.The key problems such as the selective challenge of the mass mode of the large-scale mode MIMO,the real-time optimization,the development of the 5 G real-time optimization platform for supporting the industrial Internet and the like are solved.

关 键 词：5G MASSIVE MIMO波束自配置 机器学习 MR+OTT 高精度 实时定位 

分 类 号：TN929.5[电子电信—通信与信息系统]