On the Information Content of RFID Fingerprints
Francesca M.C. Nanni, Gaetano Marrocco · 2024
In the realm of physical security methods, finger-printing stands out as a prominent measure in wireless systems and Radio-Frequency Identification (RFID) applications. The conventional approach in the literature is to extract features from the signals received from the device, and use machine learning to develop classifiers that can recognize counterfeit or cloned devices. With the goal of mastering the physical source of electromagnetic signatures, this paper investigates the information content of fingerprints of passive RFID devices by applying Shannon Information Theory to backscattered signals. The study demonstrates that by interrogating the tags with different reader's input powers and frequencies, we are able to enhance the integrated circuit (IC) non-linear behavior, and its impedance modulation will change in a not easily predictable way. Accordingly the device will expose more information, resulting in a higher entropy value, and a richer fingerprint.