Adaptive orbit recovery without navigation message and clock support for challenging satellite missions

With satellite missions increasingly operating in real-time, autonomous, or communication-limited environments, there is a critical need for orbit determination methods that remain robust without reliance on external clock corrections or broadcast navigation message. Our results reveal that the international GNSSs service standard processing often exhibits degraded accuracy or fails to converge in the orbit solution under this severe condition. To address this challenge, we present an adaptive framework for precise satellite orbit recovery that operates without requiring external clock data. We investigate a dynamic iterative least-squares adjustment driven by quality indicator, leveraging variance patterns in post-fit residuals across a network of ground tracking stations, to resolve large initial orbit uncertainties. Considering the similarities in conceptual constellation design between Korea Positioning System (KPS) and Japan’s Quasi-Zenith Satellite System (QZSS), validations are performed using real satellite-to-ground QZSS tracking observations, while large initial orbit errors ranging from several to hundreds of kilometers are introduced. Results demonstrate that the adaptive method effectively eliminates decimeter- to meter-level orbit errors resulting from the degraded initial conditions in standard orbit determination (OD) processing. The adaptive framework is particularly well suited for environments lacking precise timing or navigation messages and has therefore been incorporated into the current KPS OD platform operated within Korea Astronomy and Space Science Institute (KASI). In light of the above, this framework enables autonomous and resilient orbit recovery for a broad range of satellite missions, including GNSS constellations, remote sensing platforms, low Earth orbit tracking, and future lunar satellite navigation.