• 中文核心期刊要目总览
  • 中国科技核心期刊
  • 中国科学引文数据库(CSCD)
  • 中国科技论文与引文数据库(CSTPCD)
  • 中国学术期刊文摘数据库(CSAD)
  • 中国学术期刊(网络版)(CNKI)
  • 中文科技期刊数据库
  • 万方数据知识服务平台
  • 中国超星期刊域出版平台
  • 国家科技学术期刊开放平台
  • 荷兰文摘与引文数据库(SCOPUS)
  • 日本科学技术振兴机构数据库(JST)

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

微信公众号

一种基于LSTM与IRI模型的电离层层析TEC组合预测方法

尹萍 闫晓鹏 宁泽浩

尹萍,闫晓鹏,宁泽浩. 一种基于LSTM与IRI模型的电离层层析TEC组合预测方法[J]. 电波科学学报,xxxx,x(x): x-xx. DOI: 10.12265/j.cjors.2021271
引用本文: 尹萍,闫晓鹏,宁泽浩. 一种基于LSTM与IRI模型的电离层层析TEC组合预测方法[J]. 电波科学学报,xxxx,x(x): x-xx. DOI: 10.12265/j.cjors.2021271
YIN P, YAN X P, NING Z H. A Combined forecasting method of ionospheric tomography TEC based on LSTM and IRI model[J]. Chinese journal of radio science,xxxx,x(x): x-xx. (in Chinese). DOI: 10.12265/j.cjors.2021271
Citation: YIN P, YAN X P, NING Z H. A Combined forecasting method of ionospheric tomography TEC based on LSTM and IRI model[J]. Chinese journal of radio science,xxxx,x(x): x-xx. (in Chinese). DOI: 10.12265/j.cjors.2021271

一种基于LSTM与IRI模型的电离层层析TEC组合预测方法

doi: 10.12265/j.cjors.2021271
基金项目: 国家重点研发计划重大科学仪器设备开发专项(2018YFF01013702)
详细信息
    作者简介:

    尹萍:(1970—),女,天津人,副研究员,博士,硕士生导师,主要研究方向为卫星导航技术、电离层/大气层遥感遥测、层析成像与探测

    闫晓鹏:(1995—),男,辽宁人,硕士研究生,研究方向为卫星导航技术、电离层/大气层遥感遥测、层析成像与探测

    宁泽浩:(1997—),男,天津人,硕士研究生,研究方向为卫星导航技术、电离层/大气层遥感遥测、层析成像与探测

    通讯作者:

    闫晓鹏 E-mail: 504603278@qq.com

  • 中图分类号: P352.7

A Combined forecasting method of ionospheric tomography TEC based on LSTM and IRI model

  • 摘要: 电离层总电子含量(total electron content, TEC)作为评估无线电波穿过电离层时产生误差的主要物理量,对其准确的估算以及预测具有重要的研究意义. 本文结合电离层层析算法反演重构的TEC数据,分别采用国际参考电离层(international reference ionosphere, IRI)梯度法、长短期记忆(long short-term memory, LSTM)网络以及本文最新提出的一种基于LSTM与IRI模型的组合预测模型实现了对欧洲上空平静态电离层的TEC预测. 实验结果表明,IRI梯度法提前1 h能够产生理想的预测结果,提前2 h与3 h的预测精度明显下降. LSTM模型在提前2天的预测结果表现良好,但随着迭代预测时长的增加预测结果中出现较多异常值. 统计误差显示,本文所提出的组合预测模型相比于IRI梯度法预测性能更为稳定,对单一LSTM模型修正效果明显,消除了预测结果中大部分异常值,有效提高了单一模型的预测精度. 组合预测模型与实际层析TEC之间的预测均方根误差 (root mean squared error, RMSE)为1.10 TECu,与欧洲定轨中心提供的TEC预测RMSE为1.70 TECu.
  • 图  1  LSTM单元原理图

    Fig.  1  LSTM unit schematic diagram

    图  2  组合预测模型结构

    Fig.  2  Combined forecasting model structure

    图  3  GPS地面站分布

    Fig.  3  Distribution of GPS ground stations

    图  4  2020年10月4日12:30 UT层析TEC-map(a),IRI模型TEC-map (b),1 h预测TEC-map (c),残差DTEC-map (d),2 h预测TEC-map (e),残差DTEC-map (f),3 h预测TEC-map (g),残差DTEC-map (h)

    Fig.  4  (a)Ionospheric tomography TEC-map, (b)IRI model TEC-map, (c) 1h predicted TEC-map and (d)residual DTEC-map, (e)2h predicted TEC-map and (f)residual DTEC-map, (g)3h predicted TEC-map and (h)residual DTEC-map at 12:30 UT on October 4, 2020

    图  5  10月4日IRI梯度法不同预测时长下全局RMSE

    Fig.  5  Global RMSE under different prediction durations of IRI gradient method on 4 October

    图  6  2020年10月1日12:30 UT层析TEC-map(a),LSTM模型预测TEC-map(b),预报残差DTEC-map(c);10月2日12:30 UT实际层析TEC-map(d),LSTM模型预测TEC-map(e),预报残差DTEC-map(f);10月3日12:30 UT实际层析TEC-map(g),LSTM模型预测TEC-map(h),预报残差DTEC-map(i);10月4日12:30 UT实际层析TEC-map(j),LSTM模型预测TEC-map(k),预报残差DTEC-map(l)

    Fig.  6  Ionospheric tomography TEC-map(a), LSTM model prediction TEC-map (b) and forecast residual DTEC-map (c) at 12:30 UT on 1 October,2020; Ionospheric tomography TEC-map(d), LSTM model prediction TEC-map (e) and forecast residual DTEC-map (f) at 12:30 UT on 2 October; Ionospheric tomography TEC-map(g), LSTM model prediction TEC-map (h) and forecast residual DTEC-map (i) at 12:30 UT on 3 October; Ionospheric tomography TEC-map(j), LSTM model prediction TEC-map (k) and forecast residual DTEC-map (l) at 12:30 UT on 4 October

    图  7  2020年10月1日—4日LSTM模型全局RMSE

    Fig.  7  Global RMSE of LSTM model from 1 to 4 October, 2020

    图  8  2020年10月4日12:30 UT 层析TEC-map(a),组合预测模型预测TEC-map(b)和预报残差DTEC-map(c)

    Fig.  8  Ionospheric tomography TEC-map(a), combined forecast model prediction TEC-map (b) and forecast residual DTEC-map (c) at 12:30 UT on 4 October, 2020

    图  9  10月4日组合预测模型、IRI梯度法与LSTM模型全局RMSE

    Fig.  9  The global RMSE of the combined forecasting model, IRI gradient method and LSTM model on October 4

    表  1  10月4日IRI梯度法不同预测时长下整日平均RMSE

    Tab.  1  The average RMSE of the whole day under different prediction durations of the IRI gradient method on 4 October

    基准值评价指标/TECuτ=1 hτ=2 hτ=3 h
    实际层析TECRMSE1.291.611.91
    MAE1.171.441.72
    CODE-TECRMSE1.561.761.99
    MAE1.401.541.74
    下载: 导出CSV

    表  2  10月4日三种预测方法TEC全局误差比较

    Tab.  2  Comparison of TEC global errors of three prediction methods on 4 October

    基准值评价指标
    (TECu)
    IRI梯度法
    (1 h)
    LSTM模型组合预
    测模型
    实际层析TECRMSE1.292.111.10
    MAE1.171.650.93
    CODE-TECRMSE1.562.231.70
    MAE1.401.851.53
    下载: 导出CSV
  • [1] 袁运斌, 李敏, 霍星亮, 等. 北斗三号全球导航卫星系统全球广播电离层延迟修正模型(BDGIM)应用性能评估[J]. 测绘学报,2021,50(4):436-447.

    YUAN Y B, LI M, HUO X L, et al. Research on performance of BeiDou global broadcast ionospheric delay correction model (BDGIM) of BDS-3[J]. Acta geodaetica et cartographica sinica,2021,50(4):436-447. (in Chinese)
    [2] 秦姗兰, 王文萍, 庄文泉. 电离层延迟高阶项对GPS精密单点定位精度影响分析[J]. 全球定位系统,2017,42(5):33-38.

    QIN S L, WANG W P, ZHUANG W Q. An influence analysis of High-order ionosphere on GPS precise point positioning[J]. GNSS world of China,2017,42(5):33-38. (in Chinese)
    [3] 管斌, 孙中苗, 刘晓刚, 等. 卫星测高应用中的电离层延迟滤波处理的必要性[J]. 地球物理学进展,2020,35(6):2064-2067. DOI: 10.6038/pg2020EE0050

    GUAN B, SUN Z M, LIU X G, et al. Necessity for filtering for ionospheric delay in the applications of satellite altimetry[J]. Progress in geophysics,2020,35(6):2064-2067. (in Chinese) DOI: 10.6038/pg2020EE0050
    [4] 李帅, 李仲勤, 寇瑞雄, 等. 2019年印度尼西亚Mw7.1地震震前电离层扰动分析[J]. 导航定位学报,2020,8(6):96-102. DOI: 10.3969/j.issn.2095-4999.2020.06.014

    LI S, LI Z Q, KOU R X, et al. Analysis of ionospheric disturbance before Mw7.1 earthquake in Indonesia in 2019[J]. Journal of navigation and positioning,2020,8(6):96-102. (in Chinese) DOI: 10.3969/j.issn.2095-4999.2020.06.014
    [5] JEFFREY R, FRANKE A S J, LIU C H. Ionospheric imaging using computerized tomography[J]. Radio science,1988,23(3):299-307. DOI: 10.1029/RS023i003p00299
    [6] 邹玉华, 徐继生. 一种时变三维电离层CT算法[J]. 电波科学学报,2003,18(6):638-643. DOI: 10.3969/j.issn.1005-0388.2003.06.010

    ZOU Y H, XU J S. An algorithm of time-dependent 3 -D computerized ionospheric tomography[J]. Chinese journal of radio science,2003,18(6):638-643. (in Chinese) DOI: 10.3969/j.issn.1005-0388.2003.06.010
    [7] ZHENG D Y, YAO Y B, NIE W F, et al. A new three-dimensional computerized ionospheric tomography model based on a neural network[J]. GPS solutions, 2020, 25(1). DOI: 10.1007/s10291-020-01047-1
    [8] 尹萍, 唐卫明. 卫星导航增强系统中电离层模型的改进探讨[C]//中国卫星导航学术年会组委会. 第一届中国卫星导航学术年会论文集(中). 中国卫星导航学术年会组委会: 中国卫星导航学术年会组委会, 2010: 9.

    YIN P, TANG W M. Research on the improvement of ionospheric models applied for satellite navigation augmentation systems[C]// Organizing Committee of China satellite navigation academic annual conference. Proceedings of the first China satellite navigation Academic Annual Conference (middle). Organizing Committee of China satellite navigation academic annual conference: Organizing Committee of China satellite navigation academic annual conference, 2010: 9. (in Chinese)
    [9] 武文俊. 电离层TEC预报模型的研究[D]. 北京: 中国科学院研究生院(国家授时中心), 2008.

    WU W J. Study of ionosphere TEC prediction Models[D]. Beijing: Graduate School of Chinese Academy of Sciences (National Time Service Center), 2008. (in Chinese)
    [10] BILITZA D, REINISCH B W. International Reference Ionosphere 2007: improvements and new parameters[J]. Advances in space research,2007,42(4):599-609.
    [11] CHEN M, LIU L, XU C, et al. Improved IRI-2016 model based on BeiDou GEO TEC ingestion across China[J]. GPS Solutions,2019,24(A9):433-448.
    [12] 陈鹏, 姚宜斌, 吴寒. 利用时间序列分析预报电离层TEC[J]. 武汉大学学报(信息科学版),2011,36(3):267-270.

    CHEN P, YAO Y B, WU H. TEC Prediction of ionosphere based on time series analysis[J]. Geomatics and information science of Wuhan University,2011,36(3):267-270. (in Chinese)
    [13] 鲁铁定, 黄佳伟, 鲁春阳, 等. 基于EWT- ARMA的短期电离层TEC预测模型[J]. 大地测量与地球动力学,2021,41(4):331-335.

    LU T D, HUANG J W, LU C Y, et al. Short-term ionospheric TEC prediction model based on EWT-ARMA[J]. Journal of Geodesy and Geodynamics,2021,41(4):331-335. (in Chinese)
    [14] 袁天娇, 陈艳红, 刘四清, 等. 基于深度学习递归神经网络的电离层总电子含量经验预报模型[J]. 空间科学学报,2018,38(1):48-57. DOI: 10.11728/cjss2018.01.048

    YUAN T J, CHEN Y H, LIU S Q, et al. Prediction model for ionospheric total electron content based on deep learning recurrent neural network[J]. Chinese journal of space science,2018,38(1):48-57. (in Chinese) DOI: 10.11728/cjss2018.01.048
    [15] 汤俊, 李垠健, 钟正宇, 等. EOF-LSTM神经网络的电离层TEC预报模型[J]. 大地测量与地球动力学,2021,41(9):911-915+944.

    TANG J, LI Y J, ZHONG Z Y, et al. Prediction model of ionospheric TEC by EOF and LSTM neural network[J]. Journal of geodesy and geodynamics,2021,41(9):911-915+944. (in Chinese)
    [16] SUN W, LONG X, XIN H, et al. Forecasting of ionospheric vertical total electron content (TEC) using LSTM networks[C]// International Conference on Machine Learning and Cybernetics (ICMLC), 2017.
    [17] 马国辉, 罗云琪. 基于深度学习GRU模型的电离层总电子含量预报[J]. 测绘与空间地理信息,2020,43(S1):212-215.

    MA G H, LUO Y Q. Prediction of ionospheric TEC based on deep learning GRU model[J]. Geomatics & spatial information technology,2020,43(S1):212-215. (in Chinese)
    [18] 张富彬, 周晨, 王成, 等. 基于深度学习的全球电离层TEC预测[J]. 电波科学学报,2021,36(4):553-561.

    ZHANG F B, ZHOU C, WANG C, et al. Global ionospheric TEC prediction based on deep learning[J]. Chinese journal of radio science,2021,36(4):553-561. (in Chinese)
    [19] 廖文梯, 陈洲, 赵瑜馨, 等. 利用混合模型LSTM-DNN进行全球电离层TEC map的中短期预报[J]. 航天器环境工程,2021,38(03):281-286.

    LIAO W T, CHEN Z, ZHAO Y X, et al. Short- and medium-term forecasting of global ionospheric TEC map based on hybrid model LSTM-DNN[J]. Spacecraft environment engineering,2021,38(03):281-286. (in Chinese)
    [20] SPENCER P S J, MITCHELL C N. A three-dimensional time-dependent algorithm for ionospheric imaging using GPS[J]. Annals of geophysics,2003,46(4):687-696.
    [21] DEAR R M. Wide-area forecasting of total electron content over Europe[D]. ProQuest Dissertations and Theses Full-text Search Platform, University of Bath, United Kingdom, 2007.
    [22] DORFFNER G. Neural networks for time series processing[J]. Neural network world,1996,6:447-468.
    [23] CONNOR J T, MARTIN R D, ATLAS L E. Recurrent neural networks and robust time series prediction[J]. IEEE transactions on neural networks,2002,5(2):240-254.
    [24] HOCHREITER S. Recurrent neural net learning and vanishing gradient[J]. International Journal of Uncertainty, Fuzziness and Knowledge-Based System,1998,6(2):107-116. DOI: 10.1142/S0218488598000094
    [25] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation,1997,9(8):1735-1780. DOI: 10.1162/neco.1997.9.8.1735
    [26] 葛文, 何文峰, 陈锦林, 等. 组合模型在隧道结构沉降预测中的应用研究[J]. 地理空间信息,2021,19(8):31-34+4. DOI: 10.3969/j.issn.1672-4623.2021.08.008

    GE W, HE W F, CHEN J L, et al. Application of combined model in tunnel structure settlement prediction[J]. Geospatial information,2021,19(8):31-34+4. (in Chinese) DOI: 10.3969/j.issn.1672-4623.2021.08.008
    [27] 凌立文, 张大斌. 组合预测模型构建方法及其应用研究综述[J]. 统计与决策,2019,35(01):18-23.

    LING L W, ZHANG D B. A review of construction and application of combination forecast model[J]. Statistics & decision,2019,35(01):18-23. (in Chinese)
    [28] 鲍一丹, 吴燕萍, 何勇. BP神经网络最优组合预测方法及其应用[J]. 农机化研究,2004(3):162-164. DOI: 10.3969/j.issn.1003-188X.2004.03.073

    BAO Y D, WU Y P, HE Y. Optimal mix forecasting method based on BP neutral network and its application[J]. Journal of agricultural mechanization research,2004(3):162-164. DOI: 10.3969/j.issn.1003-188X.2004.03.073
  • 加载中
图(9) / 表(2)
计量
  • 文章访问数:  77
  • HTML全文浏览量:  22
  • PDF下载量:  10
  • 被引次数: 0
出版历程
  • 收稿日期:  2021-10-11
  • 录用日期:  2022-04-22
  • 网络出版日期:  2022-04-22

目录

    /

    返回文章
    返回