Wireless Device Identification Based on Radio Frequency Fingerprint Features

With the development of the Internet of Things (IoT) technology and the rapid deployment of 5G wireless, more and more radiation devices are appearing in the increasingly complex electromagnetic environment. To be able to manage these devices in a unified manner, accurate identification of the devices has become a top priority. Specific emitter identification (SEI) is to effectively solve this problem. In this paper, both power spectral density (PSD) and fractional Fourier transform (FrFT) methods are used to extract the characteristics of transient signals. The characteristics of steady-state signals are analyzed by the bispectrum method. The SEI system model in this paper is constructed based on these techniques. Our experiments results show that when the SNR is 16dB, the SEI system can achieve a recognition accuracy of over 97% by exploiting the characteristics of the transient signal. Since the characteristics of the steady-state signal can better suppress noise, the SEI system can achieve a nearly 90% classification recognition accuracy under extremely low SNR.