LEO Broadband Communication Satellite-Based Passive Radar Imaging: (I) Blind Signal Estimation and Structure-Aware Suppression of OFDM-Induced Ghost Responses

LEO and non-terrestrial-network (NTN) broadband communication satellites provide OFDM downlinks that can be reused as non-cooperative illuminators for passive sensing. However, communication-oriented structures such as cyclic-prefix repetition and periodically inserted synchronization resources generate deterministic correlations that appear as secondary responses in the reference-surveillance cross-ambiguity output. This paper develops a payload-blind, frame-structure-informed reference channel processing scheme for OFDM-structure-induced artifact suppression in LEO communication-signal-based passive radar imaging. Symbol- and synchronization-scale timing parameters are estimated from cyclostationary features in the reference channel within bounded receiver-side candidate sets and without pay load demodulation. The estimated structure then drives CP-plus tail editing and synchronization-component whitening in the reference channel, while the surveillance channel remains unedited. Simulations using a generic 5G-NTN-like LEO OFDM configuration show reductions of 14.3 dB and 22.4 dB for CP-related range delay artifacts and synchronization-related slow-time/Doppler frequency artifacts, respectively. A controlled reflectivity-map based passive SAR example illustrates the downstream image domain effect. A measured terrestrial LTE base-station OFDM experiment provides complementary evidence through direct path/reference-waveform processing and measured surveillance echo target-response formation. This terrestrial communication illuminator experiment evaluates the reference-processing chain with real sampled signals, hardware, noise, and propagation, but it is not a measured LEO downlink experiment or passive-SAR field validation.