欺骗攻击
阿达布思
全球导航卫星系统应用
计算机科学
人工智能
残余物
模式识别(心理学)
机器学习
支持向量机
全球定位系统
算法
计算机安全
电信
作者
Jing Li,Zhengkun Chen,Zixuan Ran,Yiyu Xu,Xiangwei Zhu,Xuelin Yuan
标识
DOI:10.1109/isas59543.2023.10164411
摘要
GNSS is widely used in various aspects of modern life, but it remains vulnerable to spoofing attacks. As a result, detecting GNSS spoofing attacks quickly and accurately has become an important research topic. Traditional spoofing detection rely on detecting changes in single-parameter features, such as signal power and signal quality, which lack adaptability. With the development of modern artificial intelligence theories, more applicable solutions for spoofing detection have emerged. A GNSS spoofing detection algorithm based on AdaBoost is proposed in the paper. Firstly, a software receiver is used to extract eight detection features, including SQM, C/N 0 , pseudo-range-Doppler consistency, pseudo-range and Doppler solution residual, and clock drift, and then the AdaBoost model is used for training and testing. The proposed method achieves highly accurate detection rates of 97% and 98% on the two public datasets, TEXBAT and OAKBAT, respectively. Moreover, it outperforms two common machine learning methods with higher detection accuracy and stable operational efficiency.
科研通智能强力驱动
Strongly Powered by AbleSci AI