文献类型：期刊文献

中文题名：一种面向基站扇区方向角估计的改进SVM算法

英文题名：An Improved SVM Algorithm for Azimuth Estimation of Base Station Sector

作者：王海[1,2];翁晨傲[1];李克[1];骆曦[1]

第一作者：王海

机构：[1]北京联合大学智慧城市学院,北京100101;[2]东南大学计算机科学与工程学院,南京211189

第一机构：北京联合大学智慧城市学院

年份：2021

卷号：47

期号：4

起止页码：120-126

中文期刊名：计算机工程

外文期刊名：Computer Engineering

收录：CSTPCD;;Scopus;北大核心:【北大核心2020】；CSCD:【CSCD_E2021_2022】；

基金：国家自然科学基金(61972040);北京联合大学人才强校优选计划项目(BPHR2018CZ05)。

语种：中文

中文关键词：软间隔支持向量机;扇区方向角;基站信息表;移动众包感知;网络测量

外文关键词：soft-margin Support Vector Machine(SVM);sector azimuth;Base Station Almanac(BSA);mobile crowdsensing;network measurement

摘要：基站扇区方向角是电信运营商进行移动网络运维的重要工参,也是基站信息表的关键要素,以人工方式为主的扇区方向角数据采集和管理方式存在成本高、数据更新不及时等问题。在分析现有方向角估计方法局限性的基础上,通过将方向角估计问题转化为基站站址约束下的最优软间隔边界求解问题,提出一种基于软间隔支持向量机的基站扇区方向角检测方法。利用移动众包感知的方式对从海量用户智能终端上采集的基站信号时空分布数据进行方向角估计,考虑到相邻扇区样本量分布差异对估计性能的影响,引入平衡C参数对检测方法进行优化。实验结果表明,相比高斯方法和径向栅格化方法,该方法的准确性和鲁棒性较高,对样本量的依赖性较低,在提高网络运维效率和智能化水平方面具有较高的应用价值。
The azimuth of base station sector is an important parameter for the operation and maintenance of mobile networks,and also a key element of Base Station Almanac(BSA).However,currently the sector azimuth data is collected and managed manually,which increases the cost and disables real-time data update.To address the limits of the existing azimuth estimation methods,this paper simplifies azimuth estimation into the searching of optimal soft-margin of cell boundary under the constraints of base station location,and on this basis proposes an azimuth estimation method for base station sector based on soft-margin Support Vector Machine(SVM).The proposed algorithm utilizes mobile crowdsensing to estimate the azimuth based on the temporal-spatial distribution data of base station signals collected from the massive smart terminals.Considering the influence of imbalanced neighboring sector sample distribution on the estimation performance,the balanced C parameter is introduced into the algorithm for optimization.The experimental results show that compared with the Gaussian method and radial rasterization method,the proposed algorithm has higher accuracy and robustness,and is also less sensitive to the size of data samples.The new algorithm offers the prospect of improving the efficiency and intelligence of mobile network operation and maintenance.