Power Spectrum Based Wi-Fi RF Fingerprint Extraction Method

Wi-Fi signals have broader spectrum and are more susceptible to multipath interference, resulting in the decline of the recognition accuracy based on radio frequency fingerprint (RFF) features. This paper proposes a novel power spectrum based Wi-Fi RF fingerprint extraction method which can eliminate the impact of multipath channels and achieve purer RFF characteristics. The experimental results show that the recognition rate of 27 devices can reach 93.3% when random forest model is used.