Перевод названия: Precise modeling of slant total electron contents with multidimensional free interpolation
Тип публикации: статья из журнала
Год издания: 2018
Идентификатор DOI: 10.18127/j20700784-201812-18
Ключевые слова: ионосферные модели, полное электронное содержание, продукты IGS, глонасс, gps, интерполяция, машинное обучение, Ionosphere modeling, total electron content, IGS Products, GLOnASS, interpolation, machine learning
Аннотация: Предложена новая интерполяционная модель ионосферных задержек, способная обеспечить точность определения наклонных ПЭС 0,02 TECU относительно соответствующих наклонных ПЭС, вычисленных с помощью современных трехмерных моделей ионосферы. Standard models of ionospheric delays widely used for practical processing of navigation signalsПоказать полностьюfrom global navigation satellite systems (GNSS) usually approximate the complicated three-dimensional electron density distribution in the Earth's ionosphere and plasmasphere with one- or two-layer distributions and have errors of order 1-8 TECU (total electron content units). Publicly available modern three-dimensional ionospheric models IRI-2016 and NeQuick2 are very complex, they are not easy to use in GNSS applications and have similar large deviations from real ionospheric data. We propose a new interpolation model of ionospheric delays which is much more precise than the one- or two-layer models but simple enough to be used in radiophysics and GNSS applications. It is based on our recent free interpolation framework used previously in finding positions of GNSS satellites from SP3 data with very high precision. The free interpolation framework is not limited to polynomial, trigonometric or spherical interpolating functions and uses a simple machine learning approach for definition of the interpolation coefficients. Our model approximates the slant TEC angular distributions obtained from modern three-dimensional ionospheric models with RMS error 0,02 TECU. The new model of ionospheric delays is universal, stable and have weak dependence on the position of the GNSS ground receiver. Experimental verification of our model was performed for two IGS stations: АМС4 (B=38,80312°, L= -104,524594°, H=1912,5 м) and PIE1 (B=34,301505°, L= -108,118927°, H = 2347,7 м) for STEC distributions calculated using both ionospheric models IRI-2016 and NeQuick2 for the years 2008 and 2017. We show that interpolation coefficients may be chosen equal for both stations and both years. Investigating deeper one may observe a small seasonal variations of the residuals of our model: RMS = 0,03 TECU for winter months and RMS = 0,05 TECU for summer months for the smallest number of parameters in our model N=7. For N=10 parameters one obtains RMS 0,02 TECU. Further research is planned to test the free interpolation ionospheric STEC model on real measurements from GNSS stations. Integration of our model and methods for differential code biases (DCBs in satellite transmitters and ground receivers) separation is obviously necessary for this test. Possible inclusion of higher order ionospheric effects is under study.
Журнал: Успехи современной радиоэлектроники
Выпуск журнала: № 12
Номера страниц: 90-94
ISSN журнала: 20700784
Место издания: Москва
Издатель: Закрытое акционерное общество Издательство Радиотехника