Delay and Doppler precompensation for LEO satellites using fused adaptive strong tracking filtering

Abstract

Low Earth orbit (LEO) satellite links are sensitive to ephemeris errors because fast geometry changes directly affect propagation delay and Doppler frequency. This paper proposes an orbit prediction-assisted precompensation scheme based on adaptive strong tracking Kalman filtering with a Keplerian constraint (ASTKF-K). Received ephemerides are modeled as intermittent observations in the Earth-centered, Earth-fixed frame. At reception epochs, Sage-Husa covariance adaptation and a strong tracking factor are used for robust updating. During gaps, the state is propagated with covariance prediction. A sliding window detector rejects anomalous ephemerides, and a local Keplerian constraint provides a feasible substitute state. The predicted state is used to compute delay and Doppler precompensation for terminal acquisition. Simulations show that ASTKF-K reduces the average position RMSE by 76.45% compared with conventional extrapolation under sparse ephemeris updates and lowers delay and Doppler residuals for a 100 s update period.