基于机器学习的阔叶林场景微蜂窝模型构建  

Microcell Attenuation Model of Broadleaf Forest Scene Based on Machine Learning

作　　者：严天峰[1,2,3] 李帅 赵亚楠 王逸轩 YAN Tian-feng;LI Shuai;ZHAO Ya-nan;WANG Yi-xuan(School of Electronic and Information Engineering,Lznzhou Jiaotong Univeristy,Lanzhou 730070,China;High-Precision Positioning Technology Compass Engineering Laboratory of Gansu Province,Lanzhou 730070,China;Radio Monitoring and Technology Center of Positioning Industry of Gansu Province,Lanzhou 730070,China)

机构地区：[1]兰州交通大学电子与信息工程学院,甘肃兰州730070 [2]甘肃省高精度北斗定位技术工程实验室,甘肃兰州730070 [3]甘肃省无线电监测及定位行业技术中心,甘肃兰州730070

出　　处：《测控技术》2019年第12期99-103,共5页Measurement & Control Technology

基　　金：甘肃省自然科学基金(1508RJZA071);兰州交通大学青年科学研究基金(2015008)

摘　　要：在信号传播过程中,传播路径的场景对信号的衰减影响很大。传统的微蜂窝模型虽然加入了对路径上建筑物的建模,但对于其他场景,如树林、湖泊、恶劣天气、人流密度等并未构建相关模型。通过从阔叶林场景下采集实测数据,并使用机器学习算法拟合数据,比较主流机器学习算法的拟合结果选择出最优模型,在视距条件下微蜂窝模型的基础上加入该场景的修正,提高了微蜂窝模型的精确度和适用范围。实验数据的拟合结果表明,基于机器学习中的决策树算法构建的阔叶林场景下的微蜂窝模型,具有较高的信号衰减预测精确度。In the process of signal propagation,the scene of the propagation path has a great influence on the attenuation of the signal.Although the traditional microcellular model has added modeling of buildings on the path,other scenarios such as forests,lakes,bad weather,and people s current densities are not considered.After collecting the measured data through the broadleaf forest scenario,the data are fitted via machine learning algorithms.Comparing the fitting results of mainstream machine learning algorithms,the optimal model is selected,and the correction of the scene is added on the basis of the microcell model under the line-of-sight condition,which improves the accuracy and application range of the microcell model.The fitting results of experimental data show that the microcellular model based on the decision tree algorithm in machine learning has a high signal attenuation prediction accuracy.

关 键 词：微蜂窝 机器学习 衰减 阔叶林 

分 类 号：TN011[电子电信—物理电子学] TP306[自动化与计算机技术—计算机系统结构]