A Physics-Informed Transformer for Atmospheric Ducting Propagation Loss Forecast

Atmospheric ducting is an atmospheric refraction effect on electromagnetic waves formed under specific weather conditions. Accurate forecasting of its propagation loss is of great importance in modern communication and radar systems. Traditional forecast methods face challenges due to the complexity of physical model inference and the incompleteness of empirical data, often resulting in low precision. This paper proposes a physics-informed Transformer method for forecasting propagation loss, utilizing the Helmholtz equation to forecast the propagation loss of atmospheric ducting. This method combines the advantages of physics-informed approaches with the efficiency of the Transformer architecture. By incorporating physics-informed conditions of atmospheric ducting propagation, the proposed method significantly improves forecast accuracy and efficiency. Experimental results demonstrate that, compared to the current state-of-the-art methods, our approach significantly enhances forecast performance under various atmospheric conditions. Performance metrics such as RMSE and training loss validate its effectiveness and application potential in forecasting atmospheric ducting propagation loss, providing a valuable reference for improving communication and navigation systems based on atmospheric ducting conditions.