当前位置: X-MOL 学术Astrophys. Space Sci. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Prediction of TEC using NavIC/GPS data with geostatistical method/forecasting capability comparison with other models
Astrophysics and Space Science ( IF 1.8 ) Pub Date : 2020-09-01 , DOI: 10.1007/s10509-020-03868-5
R. Mukesh , V. Karthikeyan , P. Soma , P. Sindhu

Total Electron Content (TEC) is used for calculation of Ionospheric delay. The Precise forecast of TEC is useful to correct the range measurements. TEC depends on the time of measurement, solar radiation (SSN & F10.7), geomagnetic index (AP & KP), season and geographic location of the user. In this paper, Cokriging geostatistical method is applied to build the Surrogate Model (COKSM) to estimate the range error based on predicted Vertical Total Electron Content (VTEC). The model is tested with pseudo-range measurements of L5 & S band data received from the operational NavIC/GPS receiver positioned at ACSCE, Bangalore, India and also using L1& L2 data of IGS network station. In order to assess the developed model, we have predicted and analyzed the ACSCE-NavIC TEC and IISC-GPS TEC and found that COKSM has predicted well for both NavIC (ACSCE) and GPS (IISC) TEC. The average RMSE of COKSM for NavIC TEC prediction is 1.4920 TECU, mean accuracy is 1.1151 TECU and average correlation coefficient is 0.9854. For GPS TEC prediction COKSM yields average RMSE of 1.1435 TECU, mean accuracy as 0.9080 TECU and average correlation coefficient of 0.9926. The average range error of NavIC and GPS are 0.2126 and 0.0938 meters. In order to estimate the performance of COKSM, it is compared with the Median model, Fourier series, NTCM-GL, SPM, LSTM and TIEGM/WEIMER models. Based on the comparison results it is observed that COKSM predicts well than other prediction models and provides the RMSE of 2.1480 TECU, correlation coefficient as 0.9810 and mean accuracy of 1.4044 TECU, also COKSM performs 4.6% better than Median model, 26.10% better than Fourier series model, 56.44% better than NTCM-GL, 56% superior than SPM model and 4.43% better than LSTM model. Apart from this, COKSM performance is assessed by comparing the forecasting results with TIEGM/WEIMER model during the St. Patrick’s storm and found that the average range error of COKSM is 1.65 m and TIEGM/Weimer model yields 2.06 m during the chosen period. These results indicate that COKSM gives better prediction results than other models and suitable for navigation applications.

中文翻译:

使用地统计方法使用 NavIC/GPS 数据预测 TEC/与其他模型的预测能力比较

总电子含量 (TEC) 用于计算电离层延迟。TEC 的精确预测有助于校正距离测量值。TEC 取决于测量时间、太阳辐射(SSN & F10.7)、地磁指数(AP & KP)、季节和用户的地理位置。在本文中,应用 Cokriging 地统计方法构建代理模型 (COKSM),以基于预测的垂直总电子含量 (VTEC) 来估计距离误差。该模型使用从位于印度班加罗尔 ACSCE 的操作 NavIC/GPS 接收器接收的 L5 和 S 波段数据的伪距测量进行测试,并且还使用 IGS 网络站的 L1 和 L2 数据。为了评估开发的模型,我们对 ACSCE-NavIC TEC 和 IISC-GPS TEC 进行了预测和分析,发现 COKSM 对 NavIC (ACSCE) 和 GPS (IISC) TEC 都进行了很好的预测。NavIC TEC 预测的 COKSM 平均 RMSE 为 1.4920 TECU,平均准确度为 1.1151 TECU,平均相关系数为 0.9854。对于 GPS TEC 预测,COKSM 产生的平均 RMSE 为 1.1435 TECU,平均精度为 0.9080 TECU,平均相关系数为 0.9926。NavIC 和 GPS 的平均距离误差分别为 0.2126 和 0.0938 米。为了估计 COKSM 的性能,将其与 Median 模型、傅立叶级数、NTCM-GL、SPM、LSTM 和 TIEGM/WEIMER 模型进行比较。根据对比结果可以看出,COKSM 的预测效果优于其他预测模型,并且提供的 RMSE 为 2.1480 TECU,相关系数为 0.9810,平均精度为 1.4044 TECU,COKSM 的性能也比 Median 模型好 4.6%,比傅立叶级数模型好 26.10%,比 NTCM-GL 好 56.44%,比 SPM 模型好 56%,比 LSTM 模型好 4.43%。除此之外,通过与圣帕特里克风暴期间的 TIEGM/WEIMER 模型的预测结果进行比较,对 COKSM 性能进行了评估,发现 COKSM 的平均距离误差为 1.65 m,而 TIEGM/Weimer 模型在所选时段内产生了 2.06 m。这些结果表明 COKSM 提供了比其他模型更好的预测结果并且适用于导航应用。通过与圣帕特里克风暴期间的 TIEGM/WEIMER 模型的预测结果进行比较来评估 COKSM 的性能,发现 COKSM 的平均距离误差为 1.65 m,而 TIEGM/Weimer 模型在所选时段内产生了 2.06 m。这些结果表明 COKSM 提供了比其他模型更好的预测结果并且适用于导航应用。通过与圣帕特里克风暴期间的 TIEGM/WEIMER 模型的预测结果进行比较来评估 COKSM 的性能,发现 COKSM 的平均距离误差为 1.65 m,而 TIEGM/Weimer 模型在所选时段内产生了 2.06 m。这些结果表明 COKSM 提供了比其他模型更好的预测结果并且适用于导航应用。
更新日期:2020-09-01
down
wechat
bug