Environment-aware Direct Position Estimation with LEO Systems

Abstract

This paper proposes an environment-aware direct position estimation (EDP) method for multi-low earth orbit (LEO) positioning in urban environments. Although LEO constellations provide strong received power, practical positioning is hindered by frequent satellite blockage and severe multipath. Moreover, the limited bandwidth of downlink reference signals reduces delay resolution, making line of sight (LoS)/non LoS (NLoS) separation unreliable and degrading conventional estimators. To address this, we exploit environmental information to predict candidate specular reflection points and reconstruct a geometry-consistent NLoS channel model. We then model dominant NLoS components as structured paths parameterized by the UE position and formulate the augmented objective function with the NLoS channel model. Simulation results show that the proposed method outperforms in terms of positioning root mean squared error (RMSE) compared to the conventional methods.