Track: Programmable Real-Time Networks and Applications |
| WiFi Threat Detector: A Deep Learning and RF Fingerprinting-Based Cybersecurity Framework for Wireless Networks |
| As smart cities and IoT ecosystems continue to rely heavily on wireless communication, the physical and MAC layers of IEEE 802.11 networks have emerged as critical vectors for cyber threats such as MAC spoofing, passive sniffing, rogue access points, and wireless-layer DDoS attacks. We introduce WiFi Threat Detector, an AI-powered, real-time cybersecurity framework that detects stealthy wireless attacks by leveraging radio frequency (RF) fingerprinting and deep learning on commodity hardware. The system comprises four modular components: offline deep learning model training, live RF feature capture, intelligent inference using hybrid neural architectures, and interactive visual analytics. Using low-cost Software-Defined Radios (SDRs) and CSI-capable network cards, it passively extracts I/Q samples, Channel State Information (CSI), and RSSI values. These are processed into spectrograms and temporal patterns, which are then analysed using a hybrid 2D CNN–LSTM model. Our approach achieves over 98.9% detection accuracy with sub-300 ms latency across spoofing, probing, rogue AP, and DDoS scenarios. It aligns with cutting-edge paradigms in SDN-Edge-IoT security, decentralized Zero Trust frameworks, RF-based UAV detection, and physical-layer side-channel analysis. This scalable and explainable framework advances proactive, AI-driven wireless security for next-generation connected environments. |
|
| Presentation Notes |
|
FATIMA-WI-FI-THREAT-DETECTOR-A-DEEP-LEARNING-RF-FINGERPRINTING-FRAMEWORK.pptx |
Real Time Communications Conference & Expo at Illinois Tech
IEEE International Conference
Follow Us
Share This
