Abstract
This letter presents a radio frequency fingerprinting identification (RFFI) protocol in wireless links with multi-antenna array transmitters. Multi-antenna systems are widely used in wireless communication systems for diversity and/or multiplexing. The mutual coupling (MC) effects arising from electromagnetic interactions between adjacent array elements can influence the RF characteristics of the transmitter and eventually the performance of the established links. In this letter, a novel RFF strategy is proposed to expand the differences in the RFF characteristics among wireless devices from the same vendor, with the goal of massively improving RFF classification accuracy in low to medium signal-to-noise ratio (SNR) channel conditions. Experimental results show that when classifying six power amplifiers (PAs) from the same vendor, 21% to 62% average classification accuracy improvement can be achieved by enlarging the RFF feature differences arising from the PA nonlinearity arising from the array coupling.
| Original language | English |
|---|---|
| Pages (from-to) | 1250-1254 |
| Number of pages | 5 |
| Journal | IEEE Communications Letters |
| Volume | 29 |
| Issue number | 6 |
| Early online date | 8 Apr 2025 |
| DOIs | |
| Publication status | Published - Jun 2025 |
Keywords
- Antenna mutual coupling
- convolution neural network (CNN)
- nonlinear memory effect
- radio frequency fingerprinting (RFF)