In this paper, we investigate the impact of wireless interferences on the physical Layer Authentication of wireless devices. The concept of physical layer authentication is to identify wireless devices from their RF emissions, which contain specific features (also called RF fingerprints) of the transmitter chain in the wireless device. This concept is also called Special Emitter Identification (SEI) or Radio Frequency-DNA (RF-DNA) and it has been researched in recent years using different techniques and machine learning algorithms. In ideal conditions, the classification accuracy presented in research literature can be often higher than 95% but it can degrade significantly in presence of non Line of Sight conditions or disturbances. The research community has investigated the impact of low Signal to Noise (SNR) ratios or fading effects on the classification performance, but the disturbances introduced by the presence of wireless interference has received little attention, even if this can be a common problem in unlicensed bands, where many different wireless standards could coexist. To address this gap, this paper presents an evaluation of emitter identification of IoT devices transmitting in unlicensed Industrial, Scientific and Medical (ISM) bands and in presence of wireless interference. We perform the classification using a Deep Learning approach based on stacked CNNs with different representations of the signal in the time, frequency and time-frequency domains. The result shows that the choice of the representation is quite significant to obtain a superior classification performance and that the best results are obtained using a representation based on the Continuous Wavelet Transform (CWT).
[1]
Shauna Revay,et al.
Deep Learning for RF Device Fingerprinting in Cognitive Communication Networks
,
2018,
IEEE Journal of Selected Topics in Signal Processing.
[2]
Gary Steri,et al.
Comparison of techniques for radiometric identification based on deep convolutional neural networks
,
2019,
Electronics Letters.
[3]
Samir Ouelha,et al.
Efficient software platform TFSAP 7.1 and Matlab package to compute Time-Frequency Distributions and related Time-Scale methods with extraction of signal characteristics
,
2017,
SoftwareX.
[4]
Walid Saad,et al.
Device Fingerprinting in Wireless Networks: Challenges and Opportunities
,
2015,
IEEE Communications Surveys & Tutorials.
[5]
Michael A. Temple,et al.
Gabor-based RF-DNA fingerprinting for classifying 802.16e WiMAX Mobile Subscribers
,
2012,
2012 International Conference on Computing, Networking and Communications (ICNC).
[6]
Boualem Boashash,et al.
Time-Frequency Signal Analysis and Processing: A Comprehensive Reference
,
2015
.
[7]
Ingrid Daubechies,et al.
The wavelet transform, time-frequency localization and signal analysis
,
1990,
IEEE Trans. Inf. Theory.
[8]
Michael A. Temple,et al.
Authorized and Rogue Device Discrimination Using Dimensionally Reduced RF-DNA Fingerprints
,
2015,
IEEE Transactions on Information Forensics and Security.
[9]
Gang Yu,et al.
General linear chirplet transform
,
2016
.