Swarm Navigation Using Signals of Opportunity from Uncooperative LEO Satellites

Methods have been developed and simulation-tested to perform swarm navigation using signals of opportunity from noncooperating Low-Earth-Orbit (LEO) satellites and cross-links between the swarm elements. This work seeks to exploit swarm capabilities in order to refine coarse initial estimates of the true orbits of the LEO satellites and thereby achieve accurate swarm navigation. The swarm consists of multiple quadrotor aircraft that can measure carrier Doppler shift from LEO satellites along with pseudorange between swarm members. Each swarm component carries an altimeter and a magnetometer too. A centralized Kalman filter estimates all swarm component states and all satellite states. Better characterization of Two-Line Element (TLE) uncertainties is important to such a system. Having better initial ephemerides for the satellites, as are available for Starlink satellites, is even more important. Achievable swarm position accuracy starting from TLEs is about 100 meters, but it improves to single-digit meters when using precise Starlink ephemerides. This work also presents a leapfrogging method in which successive members of a swarm of vehicles land and act as inertial reference stations. This not only significantly reduces the uncertainty bounds, but also allows the navigation algorithm to perform well even in the presence of strong wind, which is time-correlated (unbeknownst to the filter).