Defense Mechanisms in Autonomous Vehicles Using GPS Spoofing Detection

Defense Mechanisms in Autonomous Vehicles Using GPS Spoofing Detection

A.SAI SARANYA                                                                (223J1A4601)

V.MOURYA SRIKAR                                                       (223J1A4662)

G.AMRUTHA VARSHINI                                                (223J1A4622)

K.HARIN KOVIDH                                                           (223J1A4626)

K.MADHU                                                                           (223J1A4631)

Under the Esteemed Guidance of

Mrs. P.Sowjanya

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING (CYBER SECURITY)
RAGHU INSTITUTE OF TECHNOLOGY
(AUTONOMOUS)

ABSTRACT

Autonomous vehicles (AVs) rely heavily on Global Navigation Satellite Systems (GNSS) such as GPS for localization and navigation. However, these systems are highly vulnerable to sensor spoofing attacks, where adversaries manipulate GPS signals to mislead the vehicle’s perception of its position. Such attacks can lead to incorrect decision-making and pose serious safety risks. This work presents a simulation-based cybersecurity framework for detecting and mitigating GPS spoofing attacks in autonomous vehicles.

The proposed system integrates a kinematic consistency-based anomaly detection mechanism that validates GPS measurements against vehicle motion dynamics. By comparing the displacement derived from GPS data with the expected displacement computed from vehicle speed, the system identifies inconsistencies indicative of spoofing. Additionally, temporal anomaly detection is employed to detect abrupt positional jumps that violate physical constraints. Upon detection of a spoofing attack, the vehicle transitions into a safe state, executing controlled braking to prevent unsafe operation.

The system is implemented using the Webots simulation environment with a Python-based vehicle controller. A real-time attack scenario is simulated by injecting false GPS coordinates, enabling evaluation of detection, response, and recovery mechanisms. The results demonstrate that the proposed approach effectively identifies spoofing attacks and ensures safe system behavior through a complete defense lifecycle comprising detection, response, and recovery. This work contributes a practical and implementable cybersecurity solution for enhancing the resilience of autonomous vehicles against sensor-level attacks.