Radyo Frekansı (rf) Parmak İzi Kullanarak Cihaz Yetkilendirmesi

Loading...
Thumbnail Image

Date

2024

Journal Title

Journal ISSN

Volume Title

Publisher

Open Access Color

OpenAIRE Downloads

OpenAIRE Views

Research Projects

Organizational Units

Organizational Unit
Automotive Engineering
(2009)
Having started education in 2009, the Atılım university Department of Automotive Engineering offers an academic environment at international standards, with its education in English, a contemporary curriculum and ever-better and ever-developing laboratory opportunities. In addition to undergraduate degree education, the graduate program of multi-disciplinary mechanical engineering offers the opportunity for graduate and doctorate degree education automotive engineering. The Atılım University Automotive Engineering has been selected to be the best in Turkey in 2020 in the field of automotive engineering with studies in energy efficiency, motor performance, active/ passive automotive security and vehicle dynamics conducted in the already-existing laboratories of its own. Our graduates are employed at large-scale companies that operate in Turkey, such as Isuzu, Ford Otosan, Hattat, Honda, Hyundai, Karsan, Man, Mercedes-Benz, Otokar, Renault, Temsa, Tofaş, Toyota, Türk Traktör, Volkswagen (to start operation in 2020). In addition, our graduates have been hired at institutions such as Tübitak, Tai, Aselsan, FNSS, Ministry of National Defence, Tcdd etc. or at supplier industries in Turkey. Due to the recent evolution undergone by the automotive industry with the development of electric, hybrid and autonomous vehicle technologies, automotive engineering has gained popularity, and is becoming ever more exhilarating. In addition to combustion engine technologies, our students also gain expertise in these fields. The “Formula Student Car” contest organized since 2011 by the Society of Automotive Engineers (SAE) where our Department ranked third globally in 2016 is one of the top projects conducted by our department where we value hands-on training. Our curriculum, updated in 2020, focuses on computer calculation and simulation courses, as well as laboratory practice, catered to modern automotive technologies.

Journal Issue

Abstract

Nesnelerin İnternetinin (Nİ) genişleyen kullanım alanları, kablosuz ağlardaki güvenliğin önemini daha da artırmıştır. Kısıtlı işlem kapasitesine sahip bu cihazlarda karmaşık şifreleme yöntemleri her zaman kullanışlı değildir. Bunun sonucunda Radyo Frekanslı (RF) Parmak İzi yöntemi tanıtıldı ve başarılı sonuçlar ortaya konuldu. IoT cihazların üretim aşamalarındaki donanımsal farklılıklar kullanılarak cihazlar için bir kimlik elde edilmiştir. Bu sayede cihaz sınıflandırması ve yetkilendirmesi yapmak mümkün hale gelerek fiziksel katman güvenliğine katkı sağlanmıştır. Bu uygulamalar derin öğrenme (DÖ) ile yapılarak çok başarılı sınıflandırma doğrulukları elde edilmiştir. Ancak bu modeller, uygulama açısından, hala gelişmeye ihtiyaç duymaktadır. Bu tezde, 1 boyutlu Evrişimli Sinir Ağı (ESA) modeli ile çıkarım aşamasındaki gecikmenin düşürülmesi sunulmaktadır. 55 LoRa cihazından oluşan açık kaynak bir veri seti kullanılmıştır. Ön işleme yöntemleri olan Short Time Fourier Transform (STFT) ve Fast Fourier Transform'un (FFT) sınıflandırma doğruluğu ve çıkarım süresi bağlamında karşılaştırmaları yapılmıştır. Ek olarak, sunulan model 2 boyutlu ESA modeliyle karşılaştırılmıştır. Bu hafif model, çıkarım süresi açısından önemli iyileştirme sağlarken doğruluk açısından yalnızca çok küçük ve kabul edilebilir kayıplar gözlemlenmiştir.
With the increasing usage areas of the Internet of Things (IoT), the importance of ensuring security in wireless networks has grown. Nevertheless, power-constrained devices cannot utilize intricate encryption techniques. Later on, Radio Frequency Fin gerprinting (RFF) appeared, showcasing encouraging outcomes. A unique identity was derived from the distinct hardware variations during the production phases of IoT devices. This enabled device classification and verification functions, enhancing phys ical layer security. These applications were developed using deep learning (DL) tech niques, resulting in highly accurate classification outcomes. Nevertheless, there is still room for enhancement when it comes to putting these DL models into practice. This thesis discusses decreasing inference latency through a lightweight 1D Convolutional Neural Network (CNN) model. A dataset containing 55 LoRa devices in an open-set was utilized. Preprocessing methods of Short Time Fourier Transform (STFT) and Fast Fourier Transform (FFT) were compared based on classification accuracy and in ference latency. Furthermore, the model that was introduced was evaluated against the 2D CNN model. Even though the lightweight model offers a notable enhancement in inference speed, there are slight and tolerable reductions in its accuracy.

Description

Keywords

Elektrik ve Elektronik Mühendisliği, Sayısal iletişim, Electrical and Electronics Engineering, Digital communication

Turkish CoHE Thesis Center URL

Fields of Science

Citation

WoS Q

Scopus Q

Source

Volume

Issue

Start Page

End Page

70

Collections

Google Scholar Logo
Google Scholar™

Sustainable Development Goals

SDG data is not available