基于QPSO-LSTM模型的电离层TEC预测
doi: 10.11728/cjss2024.05.2023-0143 cstr: 32142.14.cjss2024.05.2023-0143
-
郭文韬 男, 1996年6月出生于河北省承德市, 现为桂林电子科技大学信息与通信学院硕士研究生, 主要研究方向为卫星导航电离层TEC时空演变规律及预测. E-mail: gwt2024@126.com -
孙希延 女, 山东省安丘人, 现为桂林电子科技大学研究员, 博士生导师, 广西“八桂学者”, 主要研究方向为卫星导航、卫星通信和遥感. E-mail: sunxiyan1@163.com
作者简介:
通讯作者:
Ionospheric TEC Prediction Based on QPSO-LSTM Model
-
摘要: 针对单一LSTM模型的电离层TEC短期预报存在参数调整和性能优化困难导致预测精度低的问题, 结合量子粒子群算法(Quantum Particle Swarm Optimization, QPSO)和LSTM模型, 通过量子粒子群算法自适应确定最优解, 优化LSTM模型的参数配置, 并利用该模型预测2014年和2018年共三个时段的低、中、高纬度提前5 d的电离层TEC, 对地磁活动的平静期和扰动期的电离层TEC预测精度进行实验分析. 结果表明, 经过QPSO优化的LSTM模型对TEC进行连续5 d预测时, 相对于单一LSTM模型, QPSO-LSTM模型在太阳活动低年均方根误差最多降低了0.34 TECU, 而相对精度最多提高了2.68%, 而在太阳活动高年, 低纬度地区均方根误差最多下降了0.68 TECU, 而相对精度在高纬度地区最多提高了2.36%. 从不同的角度对比分析发现, QPSO-LSTM模型的预测精度均优于单一LSTM模型.Abstract: For the ionospheric TEC short-term prediction of a single LSTM model, there are difficulties in parameter adjustment and performance Optimization, resulting in low prediction accuracy. Quantum Particle Swarm Optimization (QPSO) and LSTM model are combined. The quantum particle swarm optimization algorithm was used to determine the optimal solution, optimize the parameter configuration of the LSTM model, and use the model to predict the ionospheric TEC of low, middle and high latitudes 5 d in advance for three periods in 2014 and 2018, and analyze the prediction accuracy of the ionospheric TEC during the quiet period and disturbance period of geomagnetic activity. The experimental results show that when the LSTM model optimized by QPSO is used to predict TEC for 5 consecutive days, compared with the single LSTM model, the root-mean-square error of the QPSO-LSTM model is reduced by 0.34 TECU at most in low solar activity years, and the relative accuracy is increased by 2.68% at most in high solar activity years. The RMS error decreases by up to 0.68 TECU at low latitudes, while the relative accuracy increases by up to 2.36% at high latitudes. From different analysis angles, it is found that the prediction accuracy of QPSO-LSTM model is better than that of single LSTM model.
-
Key words:
- LSTM /
- Quantum particle swarm optimization /
- Geomagnetic activity /
- Prediction accuracy
-
图 5 平静期两种模型TEC预测值与真实值对比
Figure 5. Comparison of TEC predicted values and real values of the two models in the quiet period
图 6 扰动期两种模型TEC预测值与真实值的对比
Figure 6. Comparison of TEC predicted values and real values of the two models during the disturbance period
图 7 平静期两种模型的TEC预测结果绝对误差和相对误差直方图
Figure 7. Histogram of absolute error and relative error of TEC prediction results of two models in quiet period
图 8 扰动期两种模型的TEC预测结果绝对误差和相对误差直方图
Figure 8. Histogram of absolute error and relative error of TEC prediction results of two models in disturbance period
图 9 太阳活动高年两种模型TEC预测值与真实值对比
Figure 9. Comparison of TEC predicted values and real values of two models in high solar activity years
图 10 太阳活动高年两种模型的TEC预测结果绝对误差和相对误差直方图
Figure 10. Histogram of absolute error and relative error of TEC prediction results for two models of high solar activity year
表 1 所选位置的描述
Table 1. Description of all locations
区域编号 经纬度坐标 描述 A1 (15°N, 100°E) 低纬度 A2 (45°N, 100°E) 中纬度 A3 (75°N, 100°E) 高纬度 
下载: 导出CSV
表 2 不同参数组合的QPSO-LSTM模型预测结果
Table 2. Prediction results of QPSO-LSTM model with different parameter combinations
PopNum 10 20 40 Maxstep 10 20 40 10 20 40 10 20 40 ERMS/TECU 1.097 1.042 1.044 1.066 1.039 0.998 1.052 1.040 0.998 Pmr 95.048 95.298 95.287 95.191 95.311 95.495 95.252 95.304 95.494
下载: 导出CSV
表 3 太阳活动低年不同模型TEC预测结果的评估系数对比
Table 3. Comparison of evaluation coefficients of TEC prediction results of different models in low solar activity years
模型 评估系数 A1 A2 A3 LSTM 均方根误差/TECU 1.74 1.03 0.53 相对精度/(%) 91.43 89.40 91.50 QPSO-LSTM 均方根误差/TECU 1.40 0.80 0.46 相对精度/(%) 93.12 92.08 92.61
下载: 导出CSV
表 4 太阳活动高年不同模型TEC预测结果的评估系数对比
Table 4. Comparison of evaluation coefficients of TEC prediction results of different models in high solar activity years
模型 评估系数 A1 A2 A3 LSTM 均方根误差/TECU 4.33 2.66 2.19 相对精度/(%) 92.27 90.80 85.85 QPSO-LSTM 均方根误差/TECU 3.65 2.43 1.83 相对精度/(%) 93.47 91.43 88.21
下载: 导出CSV
-
[1] SUN W J, WU B Y, WU Z, et al. IONISE: an ionospheric observational network for irregularity and scintillation in East and Southeast Asia[J]. Journal of Geophysical Research: Space Physics, 2020, 125(8): e2020JA028055 doi: 10.1029/2020JA028055 [2] LIU L B, YANG Y Y, LE H J, et al. Unexpected regional zonal structures in low latitude ionosphere call for a high longitudinal resolution of the global ionospheric maps[J]. Remote Sensing, 2022, 14(10): 2315 doi: 10.3390/rs14102315 [3] REDDYBATTULA K D, NELAPUDI L S, MOSES M, et al. Ionospheric TEC forecasting over an Indian low latitude location using long short-term memory (LSTM) deep learning network[J]. Universe, 2022, 8(11): 562 doi: 10.3390/universe8110562 [4] RUWALI A, KUMAR A J S, PRAKASH K B, et al. Implementation of hybrid deep learning model (LSTM-CNN) for ionospheric TEC forecasting using GPS data[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 18(6): 1004-1008 doi: 10.1109/LGRS.2020.2992633 [5] 王建敏, 唐彦, 吕楠, 等. 组合模型的电离层总电子短期预报研究[J]. 测绘科学, 2022, 47(4): 34-43, 67WANG Jianmin, TANG Yan, LYU Nan, et al. Research on combined model in short term ionospheric total electron prediction[J]. Science of Surveying and Mapping, 2022, 47(4): 34-43, 67 [6] RUGVETH V S, KHATTER K. Sensitivity analysis on Gaussian quantum-behaved particle swarm optimization control parameters[J]. Soft Computing, 2023, 27(13): 8759-8774 doi: 10.1007/s00500-023-08011-4 [7] SUN J, XU W B, LIU J. Parameter selection of quantum-behaved particle swarm optimization[C]//Proceedings of the 1st International Conference on Natural Computation. Changsha: Springer, 2005: 543-552 [8] SUN J, FANG W, WU X J, et al. Quantum-behaved particle swarm optimization: analysis of individual particle behavior and parameter selection[J]. Evolutionary Computation, 2012, 20(3): 349-393 doi: 10.1162/EVCO_a_00049 [9] SUN J, WU W B, FENG B. A global search strategy of quantum-behaved particle swarm optimization[C]//Proceedings of the IEEE Conference on Cybernetics and Intelligent Systems, 2004. Singapore: IEEE, 2004 [10] 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, 341LU Tieding, HUANG Jiawei, LU Chunyang, et al. Short-term ionospheric TEC prediction model based on EWT-ARMA[J]. Journal of Geodesy and Geodynamics, 2021, 41(4): 331-335, 341 [11] 刘昂, 王宁波, 李子申, 等. 基于极大验后推估理论的全球电离层地图预报[J]. 空间科学学报, 2022, 42(6): 1089-1099 doi: 10.11728/cjss2022.06.210812085LIU Ang, WANG Ningbo, LI Zishen, et al. Prediction of global ionospheric map using the theory of maximum posterior estimation[J]. Chinese Journal of Space Science, 2022, 42(6): 1089-1099 doi: 10.11728/cjss2022.06.210812085 [12] 吴晗, 黄玲, 刘立龙, 等. 基于SSA-LSTM的短期电离层TEC组合预报模型[J]. 大地测量与地球动力学, 2022, 42(6): 626-630, 654WU Han, HUANG Ling, LIU Lilong, et al. Short term prediction model of ionospheric TEC based on SSA-LSTM[J]. Journal of Geodesy and Geodynamics, 2022, 42(6): 626-630, 654 [13] 唐丝语, 黄智. 基于因果卷积与LSTM网络的电离层总电子含量预报[J]. 空间科学学报, 2022, 42(3): 357-365 doi: 10.11728/cjss2022.03.210401042TANG Siyu, HUANG Zhi. Prediction of ionospheric total electron content based on causal convolutional and LSTM network[J]. Chinese Journal of Space Science, 2022, 42(3): 357-365 doi: 10.11728/cjss2022.03.210401042 [14] 李磊, 黎竞, 杨晨. 基于KF-LSTM组合模型的短期电离层TEC预测[J]. 大地测量与地球动力学, 2023, 43(10): 1020-1025LI Lei, LI Jing, YANG Chen. Short-term ionospheric TEC prediction based on KF-LSTM combination model[J]. Journal of Geodesy and Geodynamics, 2023, 43(10): 1020-1025 [15] NAVA B, COÏSSON P, RADICELLA S M. A new version of the NeQuick ionosphere electron density model[J]. Journal of Atmospheric and Solar-Terrestrial Physics, 2008, 70(15): 1856-1862 doi: 10.1016/j.jastp.2008.01.015 [16] XUE J K, SHEN B. A novel swarm intelligence optimization approach: sparrow search algorithm[J]. Systems Science & Control Engineering, 2020, 8(1): 22-34 [17] 汤俊, 钟正宇, 李垠健, 等. 基于SSA-Elman神经网络的电离层TEC短期预报模型[J]. 大地测量与地球动力学, 2022, 42(4): 378-383TANG Jun, ZHONG Zhengyu, LI Yinjian, et al. Short-term prediction model of ionospheric TEC based on SSA-Elman neural network[J]. Journal of Geodesy and Geodynamics, 2022, 42(4): 378-383 -
-
下载:
计量
- 文章访问数: 513
- HTML全文浏览量: 135
- PDF下载量: 67
-
被引次数:
0(来源:Crossref)
0(来源:其他)
