Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural Networks

This paper proposes a novel Deep Learning (DL)-based approach for classifying the radio-access technology (RAT) of wireless emitters. The approach improves computational efficiency and accuracy under harsh channel conditions with respect to existing approaches. Intelligent spectrum monitoring is a c...

Full description

Bibliographic Details
Main Authors: Ebtesam Almazrouei, Gabriele Gianini, Nawaf Almoosa, Ernesto Damiani
Format: Article
Language:English
Published: MDPI AG 2021-04-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/21/7/2414
_version_ 1797539330676228096
author Ebtesam Almazrouei
Gabriele Gianini
Nawaf Almoosa
Ernesto Damiani
author_facet Ebtesam Almazrouei
Gabriele Gianini
Nawaf Almoosa
Ernesto Damiani
author_sort Ebtesam Almazrouei
collection DOAJ
description This paper proposes a novel Deep Learning (DL)-based approach for classifying the radio-access technology (RAT) of wireless emitters. The approach improves computational efficiency and accuracy under harsh channel conditions with respect to existing approaches. Intelligent spectrum monitoring is a crucial enabler for emerging wireless access environments that supports sharing of (and dynamic access to) spectral resources between multiple RATs and user classes. Emitter classification enables monitoring the varying patterns of spectral occupancy across RATs, which is instrumental in optimizing spectral utilization and interference management and supporting efficient enforcement of access regulations. Existing emitter classification approaches successfully leverage convolutional neural networks (CNNs) to recognize RAT visual features in spectrograms and other time-frequency representations; however, the corresponding classification accuracy degrades severely under harsh propagation conditions, and the computational cost of CNNs may limit their adoption in resource-constrained network edge scenarios. In this work, we propose a novel emitter classification solution consisting of a Denoising Autoencoder (DAE), which feeds a CNN classifier with lower dimensionality, denoised representations of channel-corrupted spectrograms. We demonstrate—using a standard-compliant simulation of various RATs including LTE and four latest Wi-Fi standards—that in harsh channel conditions including non-line-of-sight, large scale fading, and mobility-induced Doppler shifts, our proposed solution outperforms a wide range of standalone CNNs and other machine learning models while requiring significantly less computational resources. The maximum achieved accuracy of the emitter classifier is 100%, and the average accuracy is 91% across all the propagation conditions.
first_indexed 2024-03-10T12:43:28Z
format Article
id doaj.art-5c53cdf3682c40f3b920d15319e09130
institution Directory Open Access Journal
issn 1424-8220
language English
last_indexed 2024-03-10T12:43:28Z
publishDate 2021-04-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj.art-5c53cdf3682c40f3b920d15319e091302023-11-21T13:40:28ZengMDPI AGSensors1424-82202021-04-01217241410.3390/s21072414Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural NetworksEbtesam Almazrouei0Gabriele Gianini1Nawaf Almoosa2Ernesto Damiani3Emirates ICT Innovation Centre, Khalifa University of Science and Technology, Abu Dhabi 127788, United Arab EmiratesEmirates ICT Innovation Centre, Khalifa University of Science and Technology, Abu Dhabi 127788, United Arab EmiratesEmirates ICT Innovation Centre, Khalifa University of Science and Technology, Abu Dhabi 127788, United Arab EmiratesEmirates ICT Innovation Centre, Khalifa University of Science and Technology, Abu Dhabi 127788, United Arab EmiratesThis paper proposes a novel Deep Learning (DL)-based approach for classifying the radio-access technology (RAT) of wireless emitters. The approach improves computational efficiency and accuracy under harsh channel conditions with respect to existing approaches. Intelligent spectrum monitoring is a crucial enabler for emerging wireless access environments that supports sharing of (and dynamic access to) spectral resources between multiple RATs and user classes. Emitter classification enables monitoring the varying patterns of spectral occupancy across RATs, which is instrumental in optimizing spectral utilization and interference management and supporting efficient enforcement of access regulations. Existing emitter classification approaches successfully leverage convolutional neural networks (CNNs) to recognize RAT visual features in spectrograms and other time-frequency representations; however, the corresponding classification accuracy degrades severely under harsh propagation conditions, and the computational cost of CNNs may limit their adoption in resource-constrained network edge scenarios. In this work, we propose a novel emitter classification solution consisting of a Denoising Autoencoder (DAE), which feeds a CNN classifier with lower dimensionality, denoised representations of channel-corrupted spectrograms. We demonstrate—using a standard-compliant simulation of various RATs including LTE and four latest Wi-Fi standards—that in harsh channel conditions including non-line-of-sight, large scale fading, and mobility-induced Doppler shifts, our proposed solution outperforms a wide range of standalone CNNs and other machine learning models while requiring significantly less computational resources. The maximum achieved accuracy of the emitter classifier is 100%, and the average accuracy is 91% across all the propagation conditions.https://www.mdpi.com/1424-8220/21/7/2414Deep LearningDenoising Autoencodersconvolutional neural networksclassificationIEEE Wi-Fi protocolsLTE
spellingShingle Ebtesam Almazrouei
Gabriele Gianini
Nawaf Almoosa
Ernesto Damiani
Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural Networks
Sensors
Deep Learning
Denoising Autoencoders
convolutional neural networks
classification
IEEE Wi-Fi protocols
LTE
title Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural Networks
title_full Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural Networks
title_fullStr Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural Networks
title_full_unstemmed Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural Networks
title_short Robust Computationally-Efficient Wireless Emitter Classification Using Autoencoders and Convolutional Neural Networks
title_sort robust computationally efficient wireless emitter classification using autoencoders and convolutional neural networks
topic Deep Learning
Denoising Autoencoders
convolutional neural networks
classification
IEEE Wi-Fi protocols
LTE
url https://www.mdpi.com/1424-8220/21/7/2414
work_keys_str_mv AT ebtesamalmazrouei robustcomputationallyefficientwirelessemitterclassificationusingautoencodersandconvolutionalneuralnetworks
AT gabrielegianini robustcomputationallyefficientwirelessemitterclassificationusingautoencodersandconvolutionalneuralnetworks
AT nawafalmoosa robustcomputationallyefficientwirelessemitterclassificationusingautoencodersandconvolutionalneuralnetworks
AT ernestodamiani robustcomputationallyefficientwirelessemitterclassificationusingautoencodersandconvolutionalneuralnetworks