Passive RF for Space Domain Awareness: From Theoretical Bounds and Low-Complexity Signal Processing to Field Validation in Time Synchronization

Saehan Song, Korea Aerospace Research Institute; Youeyun Jung, Korea Aerospace Research Institute; Jaedong Seong, Korea Aerospace Research Institute; Seo-Eun Lee, Korea Aerospace Research Institute; Okchul Jung, Korea Aerospace Research Institute; Girim Kwon, University of Seoul
https://doi.org/10.64861/UVYZ6492

Keywords: Passive RF, Space Domain Awareness, Data Fusion, Bayesian CRLB, Fisher Information Matrix, Short-Time CAF, Wideband Time Dilation, Orbit Determination, GNSS Common-View, Long-Baseline Synchronization, TDoA/FDoA, LEO, GEO

Abstract:

The accelerating growth of low Earth orbit (LEO) mega-constellations and debris has increased the need for scalable, all-weather space domain awareness (SDA) capabilities. Passive radio-frequency (RF) sensing, using downlink transmissions from satellites as signals of opportunity, offers a cost-effective complement to radar and optical surveillance: it requires only receive-side infrastructure, operates regardless of weather or illumination, and can track any RF-emitting object including non-cooperative targets whose orbital elements are not publicly available. This paper presents an investigation comprising two technical pillars: (1) Bayesian estimation bounds and a low-complexity wideband signal processing algorithm, and (2) a long-baseline time synchronization field demonstration. Together, these pillars establish the core signal processing and timing foundations required for multi-site passive RF surveillance of LEO objects._x000D_
_x000D_
We first derive the fundamental estimation limits for joint Time Difference of Arrival (TDoA) and Frequency Difference of Arrival (FDoA) under realistic LEO satellite channel conditions. A signal model incorporating Rician fading, unknown transmitted waveforms, and time dilation effects arising from high orbital velocities is formulated. From this model, the Fisher Information Matrix (FIM) is constructed and the Bayesian Cramer-Rao Lower Bound (CRLB) is derived, explicitly accounting for random path gain uncertainties, which are inherently unknown when passively observing non-cooperative emitters. These bounds quantitatively link system parameters, namely signal bandwidth, observation time, SNR, to achievable estimation accuracy and serve as the benchmark against which practical algorithms are evaluated. We then address the main algorithmic barrier specific to LEO: the breakdown of conventional long-integration cross ambiguity function (CAF) processing for wideband signals. At orbital velocities exceeding 7 km/s, Doppler becomes time-varying and wideband signals exhibit time-dilation distortion that smears CAF peaks across the delay and frequency domains. We develop a short-time CAF (ST-CAF) framework that partitions the observation interval into segments where narrowband behavior is preserved, aligns segment-wise CAFs using a motion-consistent drift model, and coherently recombines them to recover full processing gain. A two-stage coarse-to-fine search further reduces the two-dimensional grid burden by over an order of magnitude, enabling near-real-time feasibility. Simulations confirm that ST-CAF restores sharp ambiguity peaks in wideband conditions where conventional CAF fails, and approaches the derived CRLB in narrowband regimes, thereby closing the loop between theoretical bounds and algorithmic performance._x000D_
_x000D_
Multi-site passive RF orbit determination (OD) critically depends on inter-station time synchronization. Our previous study  quantified this dependence through a simulation-based OD sensitivity analysis, showing that orbit accuracy improves monotonically with tighter inter-station synchronization and that timing errors must be kept within several to tens of nanoseconds for multi-site passive RF to yield meaningful orbit solutions. Guided by this previously established requirement, the present work field-validates long-baseline synchronization using a GNSS common-view (CV) prototype designed for distributed passive RF stations. CV processing cancels shared satellite clock and atmospheric delay errors between sites simultaneously observing the same GNSS constellation, providing high-precision synchronization over hundreds of kilometers without atomic clocks. A custom hardware prototype with microcontroller-based high-resolution timing and 10 ns-step digital delay steering was deployed across a ~300 km baseline. Despite using low-cost commercial OCXO-based local clocks rather than expensive atomic references, the CV-based correction reduced the measured inter-site clock offset by more than half, satisfying the previously established synchronization requirement with substantial margin and confirming the practical viability of affordable multi-site timing for passive RF operations. These results — fundamental estimation bounds, a validated wideband TDoA/FDoA estimation framework, and field-proven low-cost long-baseline synchronization — establish the measurement-layer foundation for an operational multi-site passive RF SDA network._x000D_

Date of Conference: September 15-18, 2026
Track: Astrodynamics

View Paper