基于XGBoost算法的固定翼无人机空对地无线传播模型  

Air-to-ground channel wireless propagation modeling forfixed-wing UAV based on XGBoost algorithm

作　　者：谭倩 陈盛伟 周剑 杨光平 王文靖 TAN Qian;CHEN Shengwei;ZHOU Jian;YANG Guangping;WANG Wenjing(China Mobile Chengdu Institute of Research and Development 610067,China)

机构地区：[1]中国移动(成都)产业研究院,四川成都610067

出　　处：《通信与信息技术》2023年第6期102-107,共6页Communication & Information Technology

摘　　要：传统的3GPP(3rd Generation Partnership Project)无线传播模型适用于300m高度以下的场景,并不适用于中高空固定翼无人机高动态应急下的作业场景。针对中高空固定翼无人机应急场景,创新性地考虑了动态3D天线增益值进行数据仿真,构建了适用于高空动态场景下的XGBoost(EXtreme Gradient Boosting)算法的无线智能传播模型。并将其与传统的3GPP模型、空对地A2G信道模型进行了对比测试实验,结果表明XGBoost算法模型在预测准确度上最优,模型预测误差RMSE(Root Mean Square Error)小于4.99 dB,误差均值小于0.8dB,对空对地信道传播路径损耗进行准确预测具有重要意义。The traditional 3GPP wireless propagation model is applicable to scenarios below 300m altitude,and is not applicable to operational scenarios under high dynamic contingency for medium and high altitude fixed wing UAVs.Aiming at the middle and high altitude fixed wing UAV emergency scenario,the dynamic 3D antenna gain value is innovatively considered for data simulation,and the wireless intelligent propagation model of XGBoost algorithm is constructed in the high altitude dynamic scenario.The model is com⁃pared with the traditional 3GPP model and the air-to-ground model.The results show that the XGBoost algorithm model has the best prediction accuracy.The prediction error RMSE of this model is less than 4.99dB and the mean error is less than 0.8dB.It is of great sig⁃nificance to accurately predict the propagation path loss of air-to-ground channel.

关 键 词：XGBoost算法 无线传播模型 空对地 动态3D天线增益 

分 类 号：TN915.63[电子电信—通信与信息系统]