Scholarlink

<p>Autonomous Vehicles (AVs) completely depend upon global navigation satellite systems (GNSS) for accurate location information. However, GNSS signals are still at risk of various attacks, such as Global Positioning System (GPS) spoofing attacks, which lead to erroneous and misleading location information, resulting in a substantial danger to AVs and their surroundings. This study presents a lightweight and interpretable dead-reckoning-based mechanism for detecting GPS location spoofing attacks in autonomous vehicles. The proposed approach fuses onboard sensory data, including time, speed, compass, accelerometer, gyroscope, and geolocation-related measurements, to estimate vehicle position independently of GPS during the prediction stage. Spoofing is detected by evaluating the deviation between the GPS-reported location and the DR-estimated location using a threshold calibrated through empirical analysis. If the GPS location is significantly different from DR-based location estimation, it implies that the spoofing attack took place. The efficiency of the proposed method is determined with the help of the complex experimental analysis and validation in the real-time CARLA simulator, which shows the precision, recall, F1 score, and accuracy of 1.00, 0.90, 0.94, and 0.94, respectively.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Novel GPS Location Spoofing Attack Detection Mechanism for Autonomous Vehicles

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • Muhammad Arsalan,
  • Mohammad Asim Ayaz,
  • Izaz Ahmad Khan,
  • Faizan Ullah,
  • Mohsin Kamal,
  • Faheem Khan,
  • Kaznah Alshammari

摘要

Autonomous Vehicles (AVs) completely depend upon global navigation satellite systems (GNSS) for accurate location information. However, GNSS signals are still at risk of various attacks, such as Global Positioning System (GPS) spoofing attacks, which lead to erroneous and misleading location information, resulting in a substantial danger to AVs and their surroundings. This study presents a lightweight and interpretable dead-reckoning-based mechanism for detecting GPS location spoofing attacks in autonomous vehicles. The proposed approach fuses onboard sensory data, including time, speed, compass, accelerometer, gyroscope, and geolocation-related measurements, to estimate vehicle position independently of GPS during the prediction stage. Spoofing is detected by evaluating the deviation between the GPS-reported location and the DR-estimated location using a threshold calibrated through empirical analysis. If the GPS location is significantly different from DR-based location estimation, it implies that the spoofing attack took place. The efficiency of the proposed method is determined with the help of the complex experimental analysis and validation in the real-time CARLA simulator, which shows the precision, recall, F1 score, and accuracy of 1.00, 0.90, 0.94, and 0.94, respectively.