Securing Autonomous Vehicle Movement: GPS Spoofing Detection and Correction Using a Variationally Mimetic Operator Network for Improved Operational Integrity

The Global Positioning System (GPS) enables accurate vehicle positioning by processing signals transmitted from navigation satellites. Nevertheless, GPS spoofing attacks, which send counterfeit satellite signals to deceive GPS receivers, pose a significant threat to autonomous vehicle navigation by causing incorrect position estimation, route deviations, potential collisions, and violations of traffic regulations. In this manuscript, an Autonomous Vehicle Movement for GPS Spoofing Detection and Correction utilizing a Variationally Mimetic Operator Network (AVM‐GPS‐SDC‐VMON) framework is proposed to secure autonomous vehicles by accurately detecting and correcting GPS spoofing attacks. The input data created from the Car Learning to Act (CARLA) simulator is used in this study. These data are fed to VMON for detecting and classifying the GPS signals as either authentic or spoofed. The Elk Herd Optimization Algorithm (EHOA) is introduced to optimize detection factors for better precision and flexibility. The proposed approach was implemented in Python, and experiments were analyzed utilizing well‐established statistical and computational metrics to ensure reliability, robustness, and efficiency. The proposed AVM‐GPS‐SDC‐VMON achieved 99.36% accuracy, 99.6% precision, 98% detection rate (DR), and 0.02% false alarm rate (FAR) with a reduced computational time of 1.2 s compared with existing models, including Secure Autonomous Vehicles Against GPS Spoofing Attacks (SAV‐GPS‐SA‐CNN), Infrastructure‐enabled GPS Spoofing Detection and Correction (GPS‐SDC‐OCSVM), Detection and mitigation of GPS Spoofing Attacks on Phasor Measurement Units (GPS‐SA‐PMU‐RNN) methods, respectively. These results indicate that the proposed technique is highly effective for detecting GPS spoofing, significantly enhancing the security and operational integrity of autonomous vehicle navigation systems.