Trajectory-based Serving Satellite Identification with User Terminal's Field-of-View

Low-Earth-Orbit (LEO) satellite networks, such as SpaceX's Starlink, achieved global broadband Internet coverage with significantly lower latency and higher throughput than traditional satellite Internet service providers utilizing geostationary satellites. Despite the substantial advancements, the research community lacks detailed insights into the internal mechanisms of these networks. This paper presents the first systematic study of Starlink's obstruction map and serving satellite identification. Our method achieves almost unambiguous satellite identification by incorporating satellite trajectories and proposing an accurate Field-of-View (FOV) estimation approach. We validate our methodology using multiple Starlink dishes with varying alignment parameters and latitudes across different continents. We utilize Two-Line Element data to identify the available satellites within the user terminal's FOV and examine their characteristics, in comparison to those of the serving satellites. Our approach revealed a correlation between the satellite selection strategy and the user terminal to gateway latency. The findings contribute to the broader understanding of the architecture of LEO satellite networks and their impact on user experience.