Identification and Classification of Electronic Devices Using Harmonic Radar


Beatrice Perez, Timothy J. Pierson, Gregory Mazzaro, and David Kotz. Identification and Classification of Electronic Devices Using Harmonic Radar. Proceedings of the Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT), pages 248–255. IEEE, June 2023. doi:10.1109/DCOSS-IoT58021.2023.00050. ©Copyright IEEE.


Smart home electronic devices invisibly collect, process, and exchange information with each other and with remote services, often without a home occupants' knowledge or consent. These devices may be mobile or fixed and may have wireless or wired network connections. Detecting and identifying all devices present in a home is a necessary first step to control the flow of data, but there exists no universal mechanism to detect and identify all electronic devices in a space. In this paper we present ICED (Identification and Classification of Electronic Devices), a system that can (i) identify devices from a known set of devices, and (ii) detect the presence of previously unseen devices. ICED, based on harmonic radar technology, collects measurements at the first harmonic of the radar's transmit frequency. We find that the harmonic response contains enough information to infer the type of device. It works when the device has no wireless network interface, is powered off, or attempts to evade detection. We evaluate performance on a collection of 17 devices and find that by transmitting a range of frequencies we correctly identify known devices with 97.6% accuracy and identify previously unseen devices as ‘unknown’ with 69.0% balanced accuracy.

Citable with [BibTeX]

Projects: [splice]

Keywords: [authentication] [iot] [security] [sensors] [wifi]

Available from the publisher: [DOI]

Available from the author: [bib]
The publisher does not allow us to post a pdf copy; contact me if you are unable to obtain a copy from the publisher.

[Kotz research]