An improved NeQuick-G global ionospheric TEC model with a machine learning approach

The advancements in computational resources and Artificial Intelligence (AI) technological tools can predict Global Ionospheric Maps (GIMs). The prediction of GIMs certainly helps the Global Navigation Satellite System (GNSS) users with ionospheric corrections in advance. Machine Learning algorithms are prominent in predicting the complex dynamics of ionospheric space weather over a global scale through the Total Electron Content (TEC) data. The TEC predictions of NeQuick-G single-frequency GNSS Ionospheric Correction Algorithm (ICA) of the Galileo satellite navigation system are reasonable globally. The VTEC time-series data derived from NeQuick-G and CODEGIMs TEC maps data are analyzed from January 1, 2020, to December 11, 2020. An improved version of the NeQuick-G model with Center for Orbit Determination in Europe (CODE)-CODEGIM TEC data and Support Vector Regression (SVR) machine learning model is the proposed ML-aided NeQuick-G ionospheric model. This technique trains the residuals of TEC between CODEGIM and NeQuick-G models data to obtain grid residual TEC maps like CODEGIM. The ML model has aided the NeQuick-G model in reducing RMSE errors up to 4.36 TECU. The proposed research helps to estimate ionospheric corrections for the single-frequency GNSS user community.