SPACE-BASED MULTILATERATION FOR GNSS-RESILIENT SPACE TRAFFIC MANAGEMENT

The orbit determination of active Low Earth Orbit (LEO) satellites relies overwhelmingly on onboard Global Navigation Satellite System (GNSS) receivers, introducing a systemic vulnerability to jamming, spoofing, and radio frequency interference. To enhance GNSS resilience in cooperative Space Traffic Management (STM) architectures, this work investigates space-to-space multilateration based on existing Tracking, Telemetry, and Command (TT&C) downlink signals, requiring no dedicated hardware on the tracked spacecraft. The proposed concept takes advantage of nadir-pointing S-band antennas already present on operational satellites: signals from such antennas are simultaneously received by a cluster of multiple observing satellites at lower altitude equipped with pointing receivers, and Time Difference of Arrival (TDOA) and Frequency Difference of Arrival (FDOA) measurements are extracted to estimate the target's position and velocity. To understand the impact of the proposed concept on orbit determination performance, an Extended Kalman Filter (EKF) based on J_2-perturbed orbital dynamics is used which exploits TDOA/FDOA observables. It is assumed that a priori catalogue-based information is available, so that no initial orbit determination phases are needed A complete simulation framework is developed, incorporating high-fidelity orbital dynamics, a physically consistent Radio Frequency (RF) link budget with cosine-power antenna patterns, and Cramér-Rao bounded measurement noise. Three intra-cluster receiver geometries with progressively increasing baselines (30–240 km, 60–470 km, and 130–950 km) are studied for a target at 600 km and receivers at 400 km altitude. The results reveal a fundamental trade-off between geometric quality and measurement availability. The shortest baseline geometry maintains four-receiver visibility for 52% of the observation window but yields high Position Dilution of Precision (PDOP) values, resulting in an RMS position error of about 33 m under full geometric fix conditions. The most extended geometry achieves the best estimation accuracy when all four receivers are visible, with an RMS position error of 4.8 m and a median of 1.6 m, effectively neutralizing the along-track observability limitation typical of short-baseline configurations. However, full four-receiver visibility is available for only 34% of the simulation window.