基于深度学习GRU模型的电离层总电子含量预报  

TEC Prediction of Ionosphere Based on Peep Learning LSTM Model

作　　者：马国辉 罗云琪 MA Guohui;LUO Yunqi(The First Surveying and Mapping Engineering Institute of Heilongjiang,Harbin 150025,China)

机构地区：[1]黑龙江第一测绘工程院,黑龙江哈尔滨150025

出　　处：《测绘与空间地理信息》2020年第S01期212-215,共4页Geomatics & Spatial Information Technology

摘　　要：针对TEC时间序列高噪声、非平稳、包含了线性和非线性的动态序列的特性,以IGS提供的2015年时间间隔为1 h的全年电离层TEC值为序列样本数据,基于深度学习GRU模型,建立一种提前24 h的电离层TEC值预报模型,包括模型的结构设计、模型训练和预测过程的实现算法。实验结果显示,GRU模型能很好地反映电离层TEC的变化特性,在低、中、高纬度地区平均预报残差分别为2.30个TECu、1.29个TECu和2.02个TECu,预测平均相对精度分别为88.5%、87.5%和78%。The total electron content(TEC)is a representative parameter of ionosphere.For its non-linear and non-stationary characteristics,based on the serial sample data provided by IGS for the 2015 annual ionospheric TEC value of 1 h.A 24 h ahead forecasting model for ionospheric Total Electron Content(TEC)is established.A GRU model for TEC time series prediction is presented,in which the design of network structure,the procedures and algorithms of network training and forecasting are involved.Results from numerical experiments show that the predicted results are highly fitted to the actual observation data.The average prediction residuals in different latitudes are 2.3TECu,1.29TECu and 2.02TECu,respectively.The predicted average relative accuracy is 88.5%,87.5% and 78%,respectively.

关 键 词：递归神经网络 门限循环单元 电离层预报 时间序列 TEC 

分 类 号：P209[天文地球—测绘科学与技术]