基于随机森林回归的电离层幅度闪烁指数预测
Prediction of ionospheric amplitude scintillation index based on random forest regression
钟伦珑 1刘明远 1胡铁乔 1刘永玉1
作者信息
- 1. 中国民航大学 智能信号与图像处理天津市重点实验室,天津 300300
- 折叠
摘要
为满足低成本、高精度的电离层闪烁监测需求,提出一种基于随机森林回归的闪烁指数预测模型.在卫星导航接收机输出信息基础上,计算电离层结构状态参数,形成输入参数,并进行参数筛选构建训练数据集,结合专用型电离层闪烁监测接收机观测到的闪烁指数,训练生成基于随机森林回归的幅度闪烁指数预测模型.实验结果表明,与传统电离层幅度闪烁指数计算方法相比,随机森林回归模型预测得到的闪烁指数相关性更强、精度更高.
Abstract
To fulfill the requirements of economical and high-precision ionospheric scintillation monitoring,a scintillation index prediction model based on random forest regression was proposed.Parameters which reflected ionospheric state were calculated based on the output information from satellite navigation receivers.These parameters and output information were selected to form a training dataset.Combined with the scintillation index observed by the dedicated ionospheric scintillation monitoring receiver,the formed training dataset was used to train an amplitude scintillation index prediction model based on random forest regression.Simulation results show that the scintillation index predicted using the random forest regression model has stronger correlation and higher accuracy than values obtained using the traditional method.
关键词
幅度闪烁指数/随机森林回归/电离层结构状态/载噪比/电离层闪烁监测/全球导航卫星系统/预测模型Key words
amplitude scintillation index/random forest regression/ionospheric structure state/carrier-to-noise ratio/ionospheric scintillation monitoring/global navigation satellite system/prediction model引用本文复制引用
基金项目
国家重点研发计划基金项目(2020YFB0505603)
中国民航大学研究生科研创新基金项目(2021YJS081)
出版年
2024
NETL
NSTL
万方数据