Improving security of the Internet of Things via RF fingerprinting based device identification system

Security is one of the primary concerns when designing wireless networks. Along detecting user identity, it is also important to detect the devices at the hardware level. The trivial identity create-and-discard process at higher layers of the protocol stack alone is not sufficient to effectively counter security threats, such as masquerading and Sybil attacks. To counter these attacks, various radio frequency fingerprinting-based solutions are proposed for the identification of the devices. However, these approaches use expansive devices for signal capturing and rely on high sampling rates and large feature sets for analysis. In this paper, we propose a radio frequency fingerprinting-based device identification technique. The proposed technique is tested on 4G-LTE network for combined intra and inter-manufacturer device detection. It uses low-cost software defined radio to capture smartphone emissions at a lower sampling rate, using our proposed preamble threshold-based detection algorithm. The results show that our proposed technique provides classification accuracy of 95.6% at different SNR levels.