Document Details

Document Type : Thesis 
Document Title :
Hybrid Deep Learning for Signals Automatic Modulation Classification
التعلم العميق الهجين لتصنيف التضمين التلقائي لاشارات الاتصال
 
Subject : Faculty of Engineering 
Document Language : Arabic 
Abstract : Classifying signals is a crucial ability that may be applied in many situations. Before categorizing the data, it is important to break down the signal using the Fourier Transform and other long pre-process techniques that uses statistical modulation features. Due to breakthroughs in neural network topologies, algorithms, and optimization techniques, together referred to as "deep learning" (DL), we have witnessed a huge degree of change over the previous five years. Advanced deep learning algorithms can indeed be applied to the same automatic modulation classification problem and generate excellent outcomes without the need for time-consuming and manual and complex feature extraction methods. Within the recent years, Google developed a new image detecting technique called EfficientNet. In this research, a modified EfficientNet architecture was developed as part of the research on signal modulation classification. The findings are particularly outstanding at extreme signal ratios. At lower signal-to-noise ratios, we used the Transformer Block to distinguish between various signal modulation schemes. In our proposed hybrid solution, we include both the EfficientNet and the Transformer Block. It is feasible to properly classify modulated signals using the SNR band and the hybrid model. 
Supervisor : Prof. Muhammad Moinuddin 
Thesis Type : Master Thesis 
Publishing Year : 1444 AH
2022 AD 
Co-Supervisor : Prof. Ubaid M. Al-Saggaf 
Added Date : Monday, February 27, 2023 

Researchers

Researcher Name (Arabic)Researcher Name (English)Researcher TypeDr GradeEmail
هيثم خالد الشوبكيAlshoubaki, Hitham KhalidResearcherMaster 

Files

File NameTypeDescription
 49045.pdf pdf 

Back To Researches Page