DeepRadioID
Francesco Restuccia, Salvatore D’Oro, Amani Al-Shawabka, Mauro Belgiovine, Luca Angioloni, Stratis Ioannidis, Kaushik Roy Chowdhury, Tommaso Melodia · 2019
Radio fingerprinting provides a reliable and energy-efficient IoT authentication strategy by leveraging the unique hardware-level imperfections imposed on the received wireless signal by the transmitter's radio circuitry. Most of existing approaches utilize hand-tailored protocol-specific feature extraction techniques, which can identify devices operating under a pre-defined wireless protocol only. Conversely, by mapping inputs onto a very large feature space, deep learning algorithms can be trained to fingerprint large populations of devices operating under any wireless standard.