Quantum-Classical AI Secures SAR Satellites with Physical-Layer Authentication

X-band Synthetic Aperture Radar (SAR) satellites are vital for various applications, but they currently lack robust physical-layer authentication (PLA) to secure against spoofing attacks. Traditional PLA systems, often based on radio-frequency fingerprinting and classical deep learning, struggle with the unique IQ phase nonlinearities of satellite hardware and are typically limited to lower frequencies. A new research paper introduces QUASAR, a pioneering quantum-classical hybrid architecture designed to address this vulnerability. QUASAR integrates a Convolutional Neural Network (CNN) spectrogram encoder with a Variational Quantum Circuit (VQC) to perform PLA for X-band SAR signals. This innovative fusion offers significant advantages. QUASAR is remarkably data-efficient, matching classical baseline accuracy with only 10% of the training data, which is crucial given the challenges of data collection in PLA. Furthermore, it achieves higher classification accuracy than classical methods with the same data budget. Tested against replay, crafted-IQ injection, and space-borne spoofing, QUASAR successfully rejected spoofed transmissions with high percentages, establishing a new benchmark for quantum-enhanced satellite security.