جواد سرميلي

Today, the vulnerability of Global Navigation Satellite Systems (GNSS) to deliberate interference attacks, such as jamming an‎d spoofing, has become a serious security challenge. Existing methods to counter these threats are often computationally expensive, lack real-time processing capabilities, an‎d are unsuitable for implementation on resource-constrained embedded platforms. This thesis aims to overcome these challenges by presenting a novel, fast, an‎d robust approach for detecting both traditional an‎d advanced jamming attacks. The proposed research methodology is based on a hybrid an‎d optimized artificial intelligence model that utilizes a windowing technique to extract key features from the digitized Intermediate Frequency (IF) signal. The classifier model, developed based on deep learning an‎d machine learning algorithms, is specifically designed to achieve minimal computational complexity an‎d memory consumption. This optimization enables the implementation an‎d real-time execution of the algorithm on embedded hardware. eva‎luation results demonstrate the very high efficiency of the proposed system, with an accuracy of over 99% in attack detection an‎d a very low false alarm rate. A key achievement of this research is the ability to issue a pre-saturation al‎e‎rt for jamming attacks, which allows for a timely response before the receiver completely fails. The successful implementation of this model on an embedded platform an‎d the confirmation of its real-time performance indicates the efficiency an‎d practicality of this solution for real-world an‎d critical applications.

آشكارسازي , جمينگ , هشدار پيش اشباع , يادگيري ماشين تركيبي , پلتفرم‌هاي تعبيه‌شده

Detection , Jamming , Pre-Saturation al‎e‎rt , Hybrid Machine Learning , Embedded Platforms

Javad Sormayli

Dr. Mohammad Reza Mosavi