Precise real time corrections of ionospheric effects on the signals generated from Global Navigation Satellite Systems (GNSS) are important for accurate and reliable modern GNSS applications such as autonomous driving and precise farming. The aim of providing corrections can be achieved by employing an accurate model to describe the Total Electron Content (TEC) within the ionosphere. TEC is a complex and highly variable parameter (see Table). In addition, space weather events may cause strong ionospheric perturbations and are the major risk for the GNSS performance, due to a complex chain of physical processes between Sun, interplanetary magnetic field, Earth's magnetic field and ionosphere.

Developing an accurate ionosphere forecast model that also includes a space weather component is a major challenge, but an important task to achieve the desired accuracy and reliability for GNSS applications and to provide early warning information during space weather events.

The main aim of this project is to develop a model for ionospheric TEC nowcasting and forecasting by taking into account physical aspects and utilizing state-of-art machine learning techniques. A machine learning model will be trained on a vast amount of historical data, including time series of TEC data, as well as, solar and geomagnetic data from satellites and observatories. Extensive testing and validation methods will be applied in order to identify the most appropriate machine learning algorithms and combination techniques of the input data, as well as to develop a high precision ionosphere TEC forecast model.