Isolated forest-based ZigBee Device Identification Using Adaptive Filter Coefficients
Zekun Chen, Linning Peng, Hua Fu · 2022 7th International Conference on Computer and Communication Systems (ICCCS) · 2022
With the widespread use of wireless network technology, more and more attention has been paid to the Internet of things (IoT) device security. The radio frequency fingerprint-based device identification technology can effectively improve the security of wireless devices. In this paper, a novel RF fingerprint extraction method employing least mean square (LMS) filter coefficients is proposed. The work involved in this paper includes two parts: feature extraction and feature identification. The feature extraction stage includes signal synchronization, frequency offset estimation, phase offset estimation, and adaptive filtering parameter extraction. In the feature identification stage, random forest (RF) is used to verify the distinction of this fingerprint, and isolated forest (IF) is used to identify the fingerprint from unknown devices. Our experiment is conducted on 54 ZigBee devices. The experimental results show that the recognition accuracy of legal equipment and illegal equipment is 100% and 89.83% respectively in a 10dB signal-noise ratio (SNR) environment, and 98.15% and 99.55% respectively in a 30dB SNR environment.