Research on the application of deep learning in beacon light trajectory prediction under satellite platform vibration

Abstract

In satellite laser communication, the ability of communication terminals to continuously track a beacon light is crucial for maintaining communication link stability. The primary factor affecting real-time tracking is vibration of satellite platform. A neural-network-based prediction approach for beacon light positioning is introduced to mitigate the impact of vibrations on tracking accuracy. Experiments were conducted to verify the proposed approach. Compared to traditional moving average model predictions, the prediction accuracy results improved by more than 50 % with random vibrations, particularly in the mid- and low-frequency domains. This novel approach provides valuable insights for enhancing the accuracy of beacon light tracking.