Leveraging machine learning for the detection of structured interference in Global Navigation Satellite Systems.

Area of Science:

• Navigation Systems
• Signal Processing
• Machine Learning

Background:

• Radio frequency interference, particularly Global Navigation Satellite System (GNSS) spoofing, poses a significant threat to navigation and timing services.
• Sophisticated spoofing attacks, like the secure code estimation and replay (SCER) attack, challenge existing detection and mitigation techniques.
• Accurate spoofing identification is critical for safety-of-life applications, especially in aviation, where receivers may not detect counterfeit signals.

Purpose of the Study:

• To explore the efficacy of machine learning (ML) techniques in distinguishing authentic GNSS signals from spoofed ones.
• To specifically address the challenging SCER spoofing attack using the Texas Spoofing Test Battery (TEXBAT) dataset.
• To evaluate and compare the performance of logistic regression, support vector machines (SVM), K-nearest neighbors (KNN), and decision tree models.

Main Methods:

• Utilized tracking data from delay lock loop correlators as intrinsic features for ML model training.
• Trained four distinct ML models: logistic regression, SVM, KNN, and decision tree.
• Employed a random six-fold cross-validation methodology for model training and evaluation.

Main Results:

• Both logistic regression and SVM achieved a mean F1-score of 94% in detecting spoofing.
• Logistic regression demonstrated a significant time efficiency advantage, outperforming SVM by 165dB and decision trees by a factor of 3.
• Receiver operating characteristic (ROC) curve analysis further supported the performance of the evaluated models.

Conclusions:

• Logistic regression is identified as the most desirable approach for identifying SCER structured interference due to its high accuracy and superior time efficiency.
• Machine learning, particularly logistic regression, offers a robust solution for enhancing GNSS security against advanced spoofing threats.
• The findings contribute to the development of more resilient navigation systems, especially for critical aviation applications.