Inversion of the Electron Density in the Lower Ionosphere Using Artificial Intelligence

In this letter, we propose a new method that is based on the frequency-domain finite element method and artificial intelligence (FDFEM-AI) to invert the electron density in the lower ionosphere. The highly efficient FDFEM is reported to build a dataset for training and testing neural networks. The Elman neural network (ENN) is improved by optimizing the connection weights, connection thresholds, and neural numbers. The improved ENN is employed to map the relationship between the electron density and the amplitude/phase of the VLF wave in the Earth-ionosphere waveguide. The genetic algorithm is combined with the trained ENN to retrieve the optimal electron density at different altitudes and solar zenith angles using measured VLF amplitude and phase in the waveguide. Results show a good agreement between predictions using the inverted electron density and observations, which tests the effectiveness of the FDFEM-AI. This work contributes a new perspective on electron density diagnosis to find more accurate and effective methods.