Wi-Fi device identification based on multi-domain physical layer fingerprint

Abstract

Physical layer fingerprinting is a promising solution for improving the security of Wi-Fi device in Internet of Things (IoT) scenarios, as well as enhancing the usability of Wi-Fi-based applications such as user tracking, accountability, and computer forensics. The existing methods typically require the use of expensive signal analyzers to extract features, which is not conducive to practical engineering implementation and needs improvements in accuracy and speed for wireless devices with complex physical layer modulation. We propose a device fingerprinting technique that utilizes various multi-domain features extracted from the physical layer preamble. These features differ from the signal parts or feature types mostly used in previous works. Furthermore, the robustness of the selected features is evaluated in this paper. Using random forest model to identify the fingerprints, the experimental results show that the accuracy of the proposed scheme can reach 98% for 15 different types of IoT Wi-Fi devices, and 90.76% for 10 network cards with the same type of chips.