Comparison of Receiver Front-end Differences for RF Fingerprint based IoT Device Identification

RF fingerprint (RFF) based wireless device identification is an emerging technique for authentication. In this paper, we investigate RFF based Zigbee device identification with different front-end receivers. 12 CC2530 Zigbee transmitters are employed as identification targets. The USRP and HackRF software defined radio (SDR) platforms are used as high-end and low-end receivers. Zigbee device identification is evaluated from practical measurements and crossing test. Experimental results show that it is essential to obtain the target training features with high-end USRP receiver. Moreover, it is also practical to identify accessing device with low-end HackRF receiver using the trained features from high-end USRP receiver. Finally, experimental results show that both USRP and HackRF receivers could obtain 0 identification error rate in near field line-of-sight (LOS) condition. However, in the case of remote non-line-of-sight (NLOS), low-end HackRF receiver with 8-bits analog to digital convertor (ADC) suffers from serious signal to noise ratio (SNR) degradation compared to high-end USRP receiver.