Simulation Results of Spaceborne SSA Using a Comprehensive Passive Radar Model

In this paper, a high-fidelity simulation framework is developed to assess the feasibility of tracking space debris using a large low-Earth orbit (LEO) satellite constellation equipped with onboard passive radar sensors. By exploiting illumination from a distributed network of ground-based transmitters, the constellation provides consistent line-of-sight access to debris objects at higher altitudes, enabling angles-only detection and tracking. This approach yields a scalable, automated, and cost-effective architecture for next-generation space surveillance and contributes to more resilient space traffic management. A complete angles-only passive-radar orbit-determination pipeline is introduced and demonstrated. Initial orbital states are generated using a rate-aware constrained admissible-region multiple-hypothesis filter (CAR–MHF), which fuses short-arc azimuth/elevation measurements while enforcing physically consistent geometric constraints. Following initialization, a robust unscented Kalman filter (UKF) performs nonlinear state propagation and measurement fusion across the multi-sensor network. Update integrity is maintained through sigma-point Mahalanobis gating, dynamic per-measurement covariance scaling, and bounded innovations designed to suppress weak-geometry updates. A large-scale simulation is performed using realistic transmitter–receiver geometry and a heteroscedastic measurement model driven by bistatic SNR, off-boresight angle, and elevation-margin effects. The resulting performance demonstrates stable tracking behavior and multi-kilometer orbit-determination accuracy across a diverse test population, with select tracks achieving near-kilometer precision under favorable geometry. Results for measurement model validation , performance along with the tracking accuracy of the orbit determination algorithm are presented and analyzed in detail.