Projects
Federated Learning Security Against Jamming-Assisted Insider Attacks
Studies how malicious federated-learning participants can exploit wireless jamming and unfavorable channel conditions to conceal poisoned model updates. The project develops attack models and robust aggregation and detection mechanisms that incorporate wireless-side information.
Federated Learning for UAV/GNSS Spoofing Detection
Investigates secure collaborative learning for detecting GPS/GNSS spoofing in UAV and distributed wireless systems. The work examines malicious clients, label manipulation, model poisoning, and defenses that combine statistical evidence with physical-context information.
Energy-Aware Physical-Layer Security for Resource-Constrained IoT Networks
Develops adaptive physical-layer security techniques for protecting IoT wireless transmissions against eavesdropping while limiting energy consumption. The work studies channel conditions, resource allocation, artificial noise and cooperative jamming, and security-energy tradeoffs.
Machine Learning-Based RF Fingerprinting and Wireless Authentication
Investigates identification and authentication of wireless devices using physical-layer RF characteristics. The work considers reliable device recognition under changing channels, hardware variability, interference, and adversarial conditions.
Mental Health Prediction Using Machine Learning and Social Media Data
Develops machine-learning approaches for identifying indicators of depression and related mental-health conditions from social-media data. The project examines linguistic, behavioral, and temporal features, with emphasis on limited data, model generalization, and reliable prediction.
NOMA-Based Cognitive Radio Networks
Investigates NOMA-enabled cognitive radio systems that improve spectrum utilization while maintaining secure communication in the presence of untrusted users and eavesdroppers. The work includes channel-aware power allocation, cooperative jamming, reverse SIC, secrecy analysis, and optimization under different wireless channel conditions.
Machine Learning-Based Wireless Channel Prediction and Adaptive Beamforming
Uses advanced ML to predict time-varying wireless channels. Predicted channel-state information is incorporated into ZF, MRT, RZF, MMSE, and related beamforming strategies to study performance under delayed or imperfect CSI.
Security and Intelligent Resource Allocation for Advanced Wireless Networks
Develops analytical, optimization-based, and learning-assisted security techniques for NOMA, MIMO, UAV-assisted, RIS/STAR-RIS, and spectrum-sharing networks. The research focuses on secrecy performance, power allocation, cooperative communication, artificial noise, and intelligent resource optimization.