RF Fingerprint-Based Identification of IoT Devices Using Meta-ELM Classifiers

The increasing popularity of Internet of Things (IoT) devices has made them vulnerable to attackers, making customized security measures difficult due to limited resources. To address this issue, RF fingerprints, features of the physical layer of wireless signals, are being used to identify and authenticate IoT devices. This study uses low-cost software-defined radio (SDR), a single-board computer, and open-source software to capture and record RF signals. The captured signals are analyzed using the Hilbert transform to determine features for classification. Various statistical methods are applied to select the most relevant features, with the Minimum Redundancy and Maximum Relevance technique further reducing feature size. For the first time in RF fingerprinting, ELM-based Meta-ELM, Meta Multilayer ELM (Meta-MLELM), and Meta Constrained Mixed ELM (Meta-CMELM) algorithms are used as classifiers for device identification. The performance of these classifiers is evaluated based on the RF fingerprint dataset, which consist of 3752 raw signals from 7 devices. The Meta-MLELM classifier has an accuracy rate of 88%, while the Meta-ELM classifier has an accuracy rate of around 90%. The Meta-CMELM classifier performs with an accuracy rate of 92%.