QUASAR: A Quantum-Classical Neural Network for SAR Satellite Physical-Layer Authentication

Imagine you're working on securing satellite communications, which are crucial for disaster response and military operations. Currently, most systems rely on radio-frequency fingerprinting, but these methods struggle with the unique challenges posed by X-band SAR satellites, particularly at higher frequencies. They often underperform because they can't effectively capture the complex nonlinearities in the signals, leading to vulnerabilities in authentication. This is what's called physical-layer authentication (PLA), and existing solutions typically fall short in this domain. To address these issues, QUASAR combines a convolutional neural network (CNN) with a variational quantum circuit (VQC). This hybrid approach not only enhances data efficiency—requiring only 10% of the training data to achieve comparable accuracy to classical methods—but also improves classification performance under various adversarial conditions. In tests against replay attacks, crafted-IQ injections, and space-borne spoofing, QUASAR successfully rejected a significant percentage of spoofed transmissions, marking a notable step forward in the field. Compared to prior work, QUASAR's unique architecture and its ability to operate effectively with less data represent a meaningful shift in how we can secure satellite communications. For builders in this space, this means a more robust and efficient way to authenticate signals, potentially reducing the time and resources needed for data collection and improving overall security.

The integration of quantum circuits with CNNs for physical-layer authentication is a significant advancement.

The results are supported by testing against multiple adversarial scenarios, though more extensive baselines could strengthen claims.