كسري نصراله

In recent years, the rapid growth of fifth-generation an‎d beyond communication technologies has necessitated innovative approaches to enhancing spectral efficiency, reducing propagation losses, an‎d mitigating interference an‎d malicious attacks. One of the key technologies in this context is Intelligent Reflecting Surfaces (IRS), which—by utilizing a large number of programmable elements—enables intelligent control of the propagation environment an‎d adaptive beamforming. In this thesis, a beyond-line-of-sight scenario of a multi-user OFDM system is modeled, consisting of a base station (BS), several single-antenna user equipments (UEs), an‎d an IRS with phase-adjustable elements. In this model, the direct path between the users an‎d the BS is blocked, an‎d communication is established solely through the IRS. Furthermore, to examine the system’s physical-layer security, a jammer is considered that transmits an interference signal which, after reflection from the IRS, degrades the quality of the data received at the BS. The received signal at the BS comprises the superposition of user signals, noise, an‎d the jammer’s interference. To detect the presence of the jammer, an unsupervised learning approach is employed. For this purpose, received data are generated, passed through the channel, an‎d stored as a large set of clean an‎d jammer-contaminated samples. A dense-layer autoencoder neural network is then designed an‎d trained to reconstruct jammer-free signals. The reconstruction error is subsequently used as an indicator for interference detection, an‎d the decision threshold is determined based on the statistical characteristics of the clean data. Simulation results demonstrate that the proposed approach can effectively distinguish between clean an‎d jammer-contaminated signals under noisy conditions an‎d for various jammer power levels.

آشكارسازي تداخل , تشخيص ناهنجاري , صفحات هوشمند بازتابي , جمر , فراسو

Interference Detection , Anomaly Detection , Intelligent Reflecting Surface , Jammer , Uplink

Kasra Nasrollah

Dr.S.Mohammad Razavizadeh