Citation: Huang, K.; Yang, J.; Hu, P.;
Liu H. A Novel Framework for
Open-Set Authentication of Internet
of Things Using Limited Devices.
Sensors 2022, 22, 2662. https://
doi.org/10.3390/s22072662
Academic Editor: Nikos Fotiou
Received: 19 February 2022
Accepted: 29 March 2022
Published: 30 March 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the author.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
A Novel Framework for Open-Set Authentication of Internet of
Things Using Limited Devices
Keju Huang * , Junan Yang, Pengjiang Hu and Hui Liu
College of Electronic Engineering, National University of Defense Technology, Hefei 230037, China;
yangjunan@ustc.edu (J.Y.); pjhu12@nudt.edu.cn (P.H.); liuhui17c@nudt.edu.cn (H.L.)
Abstract:
The Internet of Things (IoT) is promising to transform a wide range of fields. However, the
open nature of IoT makes it exposed to cybersecurity threats, among which identity spoofing is a typ-
ical example. Physical layer authentication, which identifies IoT devices based on the physical layer
characteristics of signals, serves as an effective way to counteract identity spoofing. In this paper, we
propose a deep learning-based framework for the open-set authentication of IoT devices. Specifically,
additive angular margin softmax (AAMSoftmax) was utilized to enhance the discriminability of
learned features and a modified OpenMAX classifier was employed to adaptively identify authorized
devices and distinguish unauthorized ones. The experimental results for both simulated data and real
ADS–B (Automatic Dependent Surveillance–Broadcast) data indicate that our framework achieved
superior performance compared to current approaches, especially when the number of devices used
for training is limited.
Keywords:
Internet of Things; cybersecurity; physical layer identification; deep learning; open-set
classification
1. Introduction
The Internet of Things (IoT), which enables communication and interaction between
various devices, promises to transform a wide range of fields. IoT devices primarily transmit
information via wireless means, the open nature of which exposes the IoT to cybersecurity
threats [
1
]. One typical cybersecurity threat, identity spoofing, which refers to the action
of assuming the identity of some other device, decreases the availability of resources and
can be dangerous in critical infrastructures [
2
]. By using spoofing identities, attackers can
gain unauthorized access to internal networks and interfere with communication between
authorized devices, which threatens the security of the wireless network. Therefore, the
network administrator must identify authorized IoT devices and reject connections from
unauthorized devices (Figure 1).
To prevent identity proofing, physical layer authentication (PLA) [
3
], which is also
known as non-cryptographic device identification (NDI) [
4
], identifies IoT devices based
on the physical layer characteristics of their transmitted signals. The feasibility of PLA
is based on the fact that the electronic circuits of devices possess specific imperfections
that are determined by production and manufacturing processes. PLA is analogous to
speaker recognition [
5
] in the sense that they both concern the characteristics of components
that emit signals, regardless of the content of the signals. PLA serves as an effective tool
against identity spoofing as it identifies devices using the physical layer characteristics
of signals that stem from hardware imperfections, which cannot be counterfeited, in
theory. Compared to cryptographic approaches for authentication, PLA does not require
sophisticated key management procedures and is hard to deceive. Therefore, PLA has
received a lot of attention in the past few years.
Sensors 2022, 22, 2662. https://doi.org/10.3390/s22072662 https://www.mdpi.com/journal/sensors
