EEMD-RBF神经网络的电离层TEC预报模型  

Prediction models of ionospheric TEC by EEMD and radial basis function neural network

作　　者：刘淑琼 高鑫 李长春[1] LIU Shuqiong;GAO Xin;LI Changchun(School of Civil Engineering and Architecture, East China Jiaotong University, Nanchang 330013, China)

机构地区：[1]华东交通大学土木建筑学院,江西南昌330013

出　　处：《测绘工程》2020年第3期15-19,共5页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(41761089);江西省自然科学基金资助项目(20181BAB203027);江西省教育厅科技项目(GJJ190345)。

摘　　要：利用中、低纬度电离层总电子含量,首次建立基于集合经验模态分解与径向基函数神经网络组合模型的电离层TEC预报模型。同时,根据地磁指数的变化特征,对低纬度电离层TEC值进行磁暴日的预报建模。实验结果表明,文中提出的方法在平静日连续5 d和磁暴日连续5 d的预测上,预报效果有明显改善。In this paper,the total electron content of the ionosphere in the middle and low latitudes is used.For the first time,this paper establishes an ensemble empirical mode decomposition and Radial Basis Function Neural network combined model of ionospheric TEC prediction model.At the same time,according to the change characteristics of the geomagnetic index,the prediction model of the disturbed period for the low latitude ionosphere TEC value is performed.The experimental result shows that on the continuous 5 days prediction value on the quiet period,and on the prediction of the quiet day and the disturbed day for 5 continuous days,the prediction effect of the method proposed in this paper has improved significantly.

关 键 词：电离层 总电子含量 EEMD RBF神经网络 预报精度 

分 类 号：P228[天文地球—大地测量学与测量工程]