论文部分内容阅读
针对电离层总电子数对于提高导航定位精度的重要性,该文采用2014年国际GPS服务公布的年积日为第1~10d的电离层总电子数为原始数据,以自回归模型、灰色系统模型、BP神经网络模型两两结合的方式完成建模和预报。通过比较不同纬度、不同预测天数的实验结果分析各模型预报精度及适用范围。结果表明,利用BP神经网络模型对AR模型的预测值进行补偿,在赤道处模型拟合的相对精度达到91.32%;残差范围控制在-1.0~0.8TECU内,预报残差最优可达到0.0TECU,从而证明了该方法可以提高电离层总电子数预报精度。
In view of the importance of the total number of ionospheric electrons in improving navigation and positioning accuracy, this paper uses the data of the total ionization ion in the 1st ~ 10th days published by the International GPS Service in 2014 as the original data, and uses the autoregressive model, gray system Model, BP neural network model to complete a combination of modeling and forecasting. Through comparing the experimental results of different latitudes and different forecasting days, the accuracy of forecasting and the applicable scope of each model are analyzed. The results show that the prediction accuracy of AR model is compensated by using BP neural network model, the relative accuracy of the model fitting at the equator reaches 91.32% and the residual range is controlled within -1.0 ~ 0.8TECU. The prediction residual can reach 0.0 TECU, which proves that the method can improve the accuracy of the total number of electrons in the ionosphere.