Learning the LoS Skyline from LEO Satellite Observations for Proactive Handover

Abstract

In non-terrestrial network deployments, local obstructions may block line-of-sight satellite links before the satellite reaches the geometric elevation mask, causing abrupt and unplanned handovers. To address this limitation, this paper proposes a map-free method for learning the local LoS skyline, defined as the obstruction elevation over azimuth, from binary availability labels derived from passive satellite signal observations at the terminal. The problem is formulated as a binary classification task in the azimuth-elevation space, where the skyline is extracted as the decision boundary of the learned blockage probability surface. Two complementary estimators are investigated, namely a Gaussian Process (GP) classifier and a neural multilayer perceptron (MLP) with circular azimuth encoding and Monte Carlo Dropout uncertainty indicators. The learned obstruction surface is then combined with satellite ephemeris information through EphemerisWindow, a trajectory-level prediction method that estimates future LoS termination events before the serving link is lost. The results show that both learned estimators improve the skyline reconstruction compared with empirical bracketing and enable proactive handover preparation without requiring 3D building maps, sky cameras, or additional environmental sensing.