An AI-Driven Intrusion Detection and Protection Framework for V۲X Vehicles against GPS Spoofing and Drone Tracking, Incorporating a Cyber Data Leakage Countermeasure Approach An Adaptive Cyber Threat Filter and Defensive GPS Spoofing Approach for Connected Vehicle Security

Connected and autonomous vehicles are increasingly exposed to combined cyber-physical threats, including GPS spoofing, injection of forged messages into Vehicle-to-Everything (V۲X) networks, location data leakage, and physical tracking by hostile drones. This paper presents a simulated framework, AI-CVIDP, that integrates three complementary defensive layers within a discrete-time simulation environment: (۱) a fusion risk assessor that combines cyber, V۲X network, and physical (drone-tracking) signals into a single risk score in the ۰-۱۰۰ range; (۲) an Adaptive Cyber Threat Filter (ACTF) that uses statistical anomaly detection with a dynamic baseline and a k-sigma threshold to block unknown network nodes and filter suspicious packets; and (۳) a Defensive GPS Spoofing mechanism that feeds falsified location coordinates to external trackers (drones), reducing the accuracy of physical vehicle tracking. The framework's performance was evaluated by running headless simulations across five comparative scenarios (no attack, combined attack without defense, attack with ACTF, attack with defensive spoofing, and attack with both mechanisms) over a ۲۴۰-second simulation horizon. Experimental results show that enabling ACTF yields an average ۶۹.۷% reduction in the post-filter computed risk, while defensive GPS spoofing reduces the drone's detection rate of the vehicle from ۸۶.۵% to ۱۴.۶% during the attack window and increases the average effective tracking distance from roughly ۱۲۶ meters to over ۳۷۲ meters. Combining both mechanisms simultaneously achieves a ۷۵% reduction in post-filter risk and lowers the drone detection rate to ۵.۲%. These findings demonstrate the potential of combined cyber-physical defensive approaches, particularly active countermeasures such as location deception, to increase the resilience of connected vehicles against multi-dimensional threats.

کلیدواژه ها:

connected vehicle security ، GPS spoofing ، cyber-physical intrusion detection ، V۲X networks ، Adaptive Cyber Threat Filter (ACTF) ، defensive location deception ، drone-based tracking ، fusion risk assessment