LLM-driven ionospheric TEC intelligent perception and short-wave propagation deviation compensation

Abstract

Ionospheric total electron content (TEC) disturbance is the core factor leading to long-range high-frequency shortwave Ionospheric sounding system propagation path deviation, and traditional physical model correction methods have poor real-time performance and low accuracy under complex ionospheric conditions. Aiming at this problem, this paper proposes an intelligent ionospheric perception and solution of shortwave propagation path method based on foundational large language model (LLM). Firstly, the LLM is fine-tuned with multi-source ionospheric observation data including TEC, ionospheric layer height and solar activity parameters to realize end-to-end high-precision prediction of spatiotemporal variation of ionospheric TEC. Then, the predicted TEC disturbance parameters are embedded into the long-range high-frequency shortwave Ionospheric sounding system electromagnetic wave propagation path model to dynamically correct the refraction error of propagation. Experimental results show that compared with traditional empirical ionospheric models, the proposed method reduces the long-range high-frequency shortwave ionospheric sounding system Propagation path deviation root mean square error (RMSE) by 32.6%, and significantly improves the adaptability under ionospheric storm disturbance, which can effectively support high-precision long-range high-frequency shortwave Ionospheric sounding system.