ABSTRACT As an essential complement to terrestrial networks, low Earth orbit (LEO) satellite networks constitute a critical component of the future integrated space‐air‐ground network architecture. Owing to the high mobility of LEO satellites and the long propagation distance of satellite‐to‐ground links, the system faces significant delay jitter issues, which adversely affect communication, real‐time performance, and stability. This study proposes a temporal dynamic convolutional network model that adapts to the elevation angle by considering the orbital dynamics of LEO satellites and environmental variations in the ionosphere and troposphere along the satellite‐ground propagation path. The proposed model enables effective modeling and dynamic compensation of propagation delay jitter under different orbital altitudes and elevation angle conditions. The simulation results demonstrate that the proposed method achieves high accuracy in delay prediction and jitter suppression, providing a robust design basis and technical support for delay optimization in LEO satellite communication systems.
Gao et al. (2026) studied this question.