Channel Robust RF Fingerprint Identification Method with Narrow Band Differential Constellation Trace Figure

RF fingerprint (RFF) identification in wideband wireless communication systems is vulnerable to interference from wireless channel variations. This paper proposes a channel robust RFF identification method with narrow band differential constellation trace figure (NB-DCTF). A particular narrowband filter bank is designed to extract narrow-band RFF features from 802.11 preamble signal. The RFF features of 12 subcarriers in the short training symbol are respectively extracted and drawn into the NB-DCTF. A convolutional neural network (CNN) is then designed to identify different devices using NB-DCTF features. In the experiment, we construct a testbed including 11 mobile phones and one universal software radio peripheral (USRP) platform as the receiver. The classification accuracy in static state can reach 98.9%, and that in mobile state can reach 92.6%.