A Mock Implementation of Fused LEO GNSS

This paper presents the first experimental demonstration of fused low Earth orbit (LEO) GNSS using pseudorange measurements, achieved through a mock implementation that distributes orbit and clock models as if supplied by the constellation operator. Obtaining resilient and accurate position, navigation, and timing (PNT) from broadband LEO constellations is a topic of vigorous ongoing research. Such constellations offer key advantages over traditional GNSS: proliferated satellite coverage, high power, wideband signals, and the potential for near-zero age-of-data clock and orbit models. SpaceX’s expanding Starlink constellation and known frame structure make it an especially compelling candidate for LEO PNT. But recent analysis of its downlink frame timing and its published orbits indicates that stand-alone opportunistic use of Starlink signals (i.e., without cooperation from Starlink nor access to accurate third-party models) could not provide pseudorange-based PNT competitive in performance with traditional GNSS. By contrast, a fused LEO GNSS approach in which Starlink supplies real-time clock and orbit corrections to subscribers, offers a promising pathway to accurate PNT despite frame timing irregularities. This study develops downlink-beam-specific parameterized orbit and clock models tailored to Starlink’s frame clock variations. Then, as if from the constellation operator in a hypothetical fused LEO GNSS service, these models are distributed to users who combine them with live-signal Starlink observables to produce navigation and timing solutions with 10-meter-level accuracy. These results demonstrate that fused LEO GNSS is a viable alternative or complement to traditional GNSS for moderately accurate and resilient PNT.