Index Modulation Leveraging RF Fingerprints: System Design and Prototype Validation

Abstract

This letter exploits radio frequency (RF) fingerprints as a new information-bearing index domain beyond their conventional use in device identification or authentication, and proposes two index-only schemes: RF fingerprint modulation (RFFM), which conveys index bits via individual RF-chain fingerprints, and generalized RFFM (GRFFM), which conveys index bits via composite RF fingerprints while supporting expansion of the index-state set. With a common antenna and a fixed baseband waveform, all index states are modeled as sharing a common frame-wise channel effect, enabling deep-learning-based RF fingerprint index recovery without explicit channel estimation. A hardware prototype is implemented, and over-the-air experiments demonstrate reliable index recovery for both schemes, with the average bit error rate reaching the $10^{-4}$ level under high-SNR conditions. Degree of cyclostationarity analysis further shows that RFFM and GRFFM exhibit weaker nonzero-cycle features, indicating reduced cyclostationary observability.