Implementation of Hybrid Ionospheric TEC Forecasting Algorithm Using PCA-NN Method

• Published in: IEEE journal of selected topics in applied earth observations and remote sensing 2019-01, Vol.12 (1), p.371-381
• Main Authors: Mallika, I. Lakshmi, Ratnam, D. Venkata, Ostuka, Yuichi, Sivavaraprasad, G., Raman, Saravana
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial intelligence; Artificial neural networks; Earth; Environmental conditions; Forecasting; Geomagnetism; Global navigation satellite system; Global navigation satellite systems (GNSS); Indexes; Ionosphere; Lithosphere; Magnetospheres; Mesosphere; Meteorology; Navigation; Navigation satellites; Neural networks; neural networks (NNs); Predictive models; Principal component analysis; principal component analysis (PCA); Principal components analysis; Satellites; space weather; Stratosphere; TEC; Temporal variations; Thermosphere; Time lag; Total Electron Content; Troposphere; Weather forecasting
• ISSN: 1939-1404; 2151-1535
• DOI: 10.1109/JSTARS.2018.2877445

Forecasting the ionospheric space weather is crucial for improving the accuracy of the global navigation satellite systems (GNSS). Nonetheless, comprehending the nonhomogeneous ionospheric variability under space earth environmental conditions is a major challenge, and so is developing an accurate ionospheric forecasting model. The complex spatial and temporal variations in the ionospheric region are the results of the solar and interplanetary activities, in addition to the magnetosphere, mesosphere, thermosphere, stratosphere, troposphere, and lithosphere processes. Thus, this calls for an urgent need to develop a suitable ionospheric forecasting algorithm to capture the ionospheric perturbations. Total electron content (TEC) is the key parameter derived from GNSS receivers to represent the status of the ionosphere. This paper introduces a novel ionospheric forecasting algorithm based on the fusion of principal component analysis and artificial neural networks (PCA-NN) methods to forecast the ionospheric TEC values. Solar index (F10.7), geomagnetic index (Ap index), and 20-year TEC data (1997-2016) over a Japan Grid Point (34.95 °N and 134.05 °E) were used to apply artificial intelligence methodologies. The experimental results underscore the reliability of the proposed algorithm in forecasting the ionospheric time delay effects.