Enhancing RFID Antenna Electromagnetic Fingerprints Through Non-Linear Interrogation

Abstract

Fingerprinting stands as an effective non-intrusive and non-destructive method to ensure physical security in wireless systems and Radio-Frequency Identification (RFID) applications. Conventionally, the most common state of the art approach involves extracting signal features from the devices and employing machine learning techniques for the classification of counterfeit or cloned ones. This paper explores how to enhance RFID antenna electromagnetic fingerprints by proposing a multi-power interrogation approach. Unlike traditional methods, our technique emphasizes the non-linear behavior of RFID integrated circuits (ICs) by properly varying the reader input power and frequencies. This strategy increases the unpredictability of the IC impedance modulation, thereby extracting richer and more complex information from the RFID tags. Using Shannon Information Theory, we can quantify the entropy of these enhanced fingerprints, revealing an average increase of almost 2 bits in the information content compared to single-power level interrogations. Our findings can lay the foundations to implement more robust RF physical unclonable functions (PUFs) with robust physical keys against counterfeiting and replication threats.