Enhancing Autonomous Vehicle Navigation in GPS‐Spoofed Environments Using Quantum Self‐Attention Neural Networks for Robust Positioning and Path Planning

ABSTRACT Autonomous vehicles (AVs) primarily depend on GPS for their location and navigation. However, GPS spoofing, where fake signals fool the receiver, poses a severe threat with incorrect localization, unsafe maneuvers, and navigation failures. This paper proposes a new approach: enhancing autonomous vehicle navigation in GPS‐spoofed environments using quantum self‐attention neural networks for robust positioning and path planning (EAVN‐GPSSE‐QSANN‐RPPP). The proposed method uses a GPS spoofing dataset. It introduces the usage of a regularized bias‐aware ensemble Kalman filter (RBEKF) for noise reduction and bias correction, a lotus effect optimizer (LEO) for selecting discriminative features, and a quantum self‐attention neural network (QSANN) optimized with the Parrot Optimizer Algorithm for an accurate spoofing detection and classification task. The proposed EAVN‐GPSSE‐QSANN‐RPPP approach attains 7.14%, 6.02%, and 8.27% higher accuracy and 7.36%, 5.48%, and 8.27% higher precision compared with existing techniques, respectively. This work confirms that the proposed architecture should be able to guarantee robust localization, improved path planning, and resilience against GPS spoofing, enhancing safety and reliability for the operation of AV navigation in adversarial environments.