Ionospheric Modeling by Using Self-Organizing Map (SOM) Under the Disturbed Condition

Ionospheric delay is the main error source of Global Navigation Satellite Systems (GNSS), and it is important to estimate an accurate understanding of ionosphere delay conditions through Total Electron Content (TEC), especially during disturbances. It is desirable that the ionospheric modeling method is as simple, responsive as possible, robust, and highly accurate. To address this challenge, in this paper, we proposed ionospheric modeling by using a Self-Organizing Map (SOM) (Kohonen., 1982) which is one of unsupervised machine learning. SOM is a method that maps from a high-dimensional space to a low-dimensional space while preserving the topology, and this method is very suitable for modeling complex distributed data in space. Since SOM has its origins in biological algorism which has topological representation self-organizing through “competitive” and “cooperative” learning processes, even complex phenomena can be modeled with flexibility and stability. In this paper, by using this proposed system, the ionosphere was modeled with only the position (Latitude and Longitude) and STEC (Slant Total Electron Content) data which is calculated by ground receivers of the GNSS Earth Observation Network System in Japan. This new approach could be applied to highly precise satellite positioning systems such as PPP-RTK including Centimeter Level Augmentation Service (QZSS CLAS) (Cabinet Office, Government of Japan., 2022), and also considering space weather, or disaster prevention in the future. Keywords: Self-Organizing Map (SOM), Ionospheric modeling