Securing Public Wi-Fi Networks Through SIM-Based Defense

Abstract

Public Wi-Fi networks have an open and shared communication environment, they are highly vulnerable to cyberspace attacks like spoofing, man-in-the-middle (MITM) attacks, intrusions, and eavesdropping. This work proposes a simulation-based security model that would improve user security and trust building of the public wireless ecosystem by combining Machine Learning (ML) and Stacked Intelligent Metasurfaces (SIM). A real-time simulation framework is developed to monitor and process the wireless traffic based on tools like Tshark and Scapy. Extracted features of networked devices are managed to determine communication habits and behavioral traits, and these are classified using trained ML models, like a Random Forest, to identify an abnormal or malicious activity. Depending on the classification results, a dynamic trust score engine determines the trustworthiness of connected devices and activates adaptive responses of SIM-based. The beamforming techniques are used to support trusted users, simulated signal degradation is applied to untrusted devices, and artificial noise injections are used to disrupt suspicious nodes. This dynamic physical layer security model can help in real-time alleviation of threats and at the same time maintain the performance of the network to authorized users. According to the results of the conducted experiment, the proposed framework can help improve the detection accuracy, make wireless security more resilient, and provide a scalable approach to the protection of the public Wi-Fi infrastructures.