Error analysis and parameter optimization of ionospheric autocorrelation prediction method
-
摘要: 电离层延迟是全球卫星导航系统(GNSS)高精度导航定位应用中的重要误差源. 通过对电离层总电子含量(TEC)进行测量和短期预报可有效提升GNSS单频用户的定位精度，对其他无线电系统的电离层效应也可起到有效减缓作用. 近二十年来提出了很多行之有效的短期预报方法，但还没有哪一种方法有绝对优势，其预测精度都有待提高. 利用Madrigal数据库任意选取的5个格网点的TEC观测数据，首先比较了自相关和自回归滑动平均(ARIMA)方法，然后研究了自相关预报方法中实际参与加权的观测值覆盖天数和观测值数量这两个参数的选取对预报误差的影响，并提出了参数设置优化方案. 试验结果显示：1)自相关方法的预报误差略小于ARIMA方法，且自相关方法所花费时间比ARIMA方法少，总体上自相关方法是一种性能更优的方法；2)对于自相关方法，相比于传统的“4+12”参数设置方案，“3+9”方案总体上具有更优的预报性能，说明TEC时间序列的当前状态可能主要与前3天的状态有关. 相关结果可作为电离层短期预报工程实现的一个有用的参考方案.Abstract: Ionospheric delay is an important error source in GNSS high-precision navigation and positioning applications. Through the measurement and short-term prediction of total ionospheric electron content (TEC), the positioning accuracy of GNSS single frequency users can be effectively improved, and the ionospheric effect of other radio systems can also be effectively alleviated. In the past two decades, many effective short-term prediction methods have been proposed, but none of them is absolutely leading, and the prediction accuracy of all these methods needs to be improved. In this paper, using the TEC observation data of five grid points arbitrarily selected from the Madrigal database, the autocorrelation and autoregressive moving average (ARIMA) methods are compared, and then the influence of two parameters on the prediction error in the autocorrelation prediction method is studied. Finally, an optimized parameter setting scheme is put forward for the autocorrelation prediction method. The experimental results show that: 1) The prediction error of the autocorrelation method is slightly smaller than that of the ARIMA method, and the time taken by the autocorrelation method is obviously less than that of the ARIMA method. Therefore, the comprehensive performance of the autocorrelation method is better than ARIMA method; 2) For the autocorrelation method, compared with the traditional “4+12”scheme, “3+9” scheme has better prediction performance on the whole, indicating that the ionosphere current state may be mainly related to the state of the previous three days. The relevant results can be used as a useful reference scheme for the implementation of ionospheric short-term prediction engineering.“3+9” scheme has better prediction performance on the whole, indicating that the ionosphere current state may be mainly related to the state of the previous three days. The relevant results can be used as a useful reference scheme for the implementation of ionospheric short-term prediction engineering.
-
Key words:
- ionosphere /
- short-time prediction /
- autocorrelation /
- error analysis /
- parameter optimization
-
表 1 不同观测站两种方案预测误差综合比较
观测站
序号RMSE/TECU NRMSE/% “4+12” “3+9” “4+12” “3+9” 1 0.535 0.521 10.10 9.82 2 0.676 0.667 9.38 9.25 3 0.720 0.690 11.48 11.02 4 1.197 1.179 13.75 13.55 5 0.999 0.967 14.07 13.62 
下载: 导出CSV
-
[1] MUHTAROV P, KUTIEV I. Autocorrelation method for temporal interpolation and short-term prediction of ionospheric data[J]. Radio science, 1999, 34(2): 459-464. DOI: 10.1029/1998RS900020 [2] MARIN D, MIRO G, MIKHAILOV A V. A method for foF2 short-term prediction[J]. Physics and chemistry of the earth, 2000, 25(4): 327-332. DOI: 10.1016/S1464-1917(00)00026-X [3] CANDER L R, MILOSAVLJEVIC M M, STANKOVIC S S, et al. Ionospheric forecasting technique by artificial neural network[J]. Electron letters, 1998, 34(16): 1573-1574. DOI: 10.1049/el:19981113 [4] LIU R Y, XU Z, WU J, et al. Preliminary studies on ionospheric forecasting in China and its surrounding area[J]. Journal of atmospheric and solar-terrestrial physics, 2005, 67(12): 1129-1136. DOI: 10.1016/j.jastp.2004.12.012 [5] LIU R Y, LIU S L, XU Z H, et al. Application of autocorrelation method on ionospheric short-term forecasting in China[J]. Chinese science bulletin, 2006, 51(3): 352-357. DOI: 10.1007/s11434-006-0352-9 [6] KOUTROUMBAS K, TSAGOURI I, BELEHAKI A. Time series autoregression technique implemented on-line in DIAS system for ionospheric forecast over Europe[J]. Annales geophysicae, 2008, 26(2): 371-386. DOI: 10.5194/angeo-26-371-2008 [7] TSAGOURI I, KOUTROUMBAS K, BELEHAKI A. Ionospheric foF2 forecast over Europe based on an autoregressive modeling technique driven by solar wind parameters[J]. Radio science, 2009, 44(1): RS0A35. DOI: 10.1029/2008RS004112 [8] TSAI L C, MACALALAD E P, LIU C H. TaiWan ionospheric model (TWIM) prediction based on time series autoregressive analysis[J]. Radio science, 2014, 49(10): 977-986. DOI: 10.1002/2014RS005448 [9] TSAGOURI I, BELEHAKI A. An upgrade of the solar-wind-driven empirical model for the middle latitude ionospheric storm-time response [J]. Journal of atmospheric and solar-terrestrial physics, 2008, 70(16): 2061-2076. DOI: 10.1016/j.jastp.2008.09.010 [10] ZHAO X K, NING B Q, LIU L B, et al. A prediction model of short-term ionospheric foF2 based on AdaBoost[J]. Advances in space research, 2014, 53(3): 387-394. DOI: 10.1016/j.asr.2013.12.001 [11] WANG J, FENG F, MA J G. An adaptive forecasting method for ionospheric critical frequency of F2 Layer[J]. Radio science, 2020, 55(1): 1-12. DOI: 10.1029/2019RS007001 [12] RIDEOUR W, COSTER A J. Automated GPS processing for global total electron content data[J]. GPS solutions, 2006, 10(3): 219-228. DOI: 10.1007/s10291-006-0029-5 -
点击查看大图
图(5) / 表(1)
计量
- 文章访问数: 169
- HTML全文浏览量: 87
- PDF下载量: 21
- 被引次数: 0
