Saturday 01 March 2025
Researchers have long been concerned about the vulnerability of Global Navigation Satellite Systems (GNSS) to spoofing and jamming attacks, which can disrupt critical infrastructure like air traffic control systems and financial transactions. To combat these threats, a team of scientists has developed advanced machine learning and deep learning techniques to detect and classify GNSS signals.
The researchers began by creating two datasets: one for spoofing detection and another for jamming signal classification. The spoofing dataset consisted of 510,530 samples, each with 13 numerical features, while the jamming dataset contained 120,000 images categorized into six classes. To address the imbalance in the datasets, the team implemented various techniques such as random oversampling, random undersampling, and synthetic minority over-sampling technique (SMOTE).
For spoofing detection, the researchers employed traditional machine learning algorithms like logistic regression, k-nearest neighbors, Gaussian Naïve Bayes, support vector machines, decision tree classifier, random forest classifier, and gradient boosting machine. They also designed custom convolutional neural networks (CNNs) to leverage the powerful representation capabilities of deep learning architectures.
In their experiments, the team achieved impressive results, with a best-performing model achieving an accuracy of approximately 94.44% in spoofing detection. For jamming signal classification, they trained a simple CNN model that incorporated residual blocks and obtained an accuracy of 97.78%. By fine-tuning the model using transfer learning, they were able to improve its performance even further.
The researchers also explored ensemble methods, combining multiple models to create more robust detectors. They found that these ensembles outperformed individual models in both spoofing detection and jamming signal classification tasks. The team’s approach is significant because it demonstrates the potential of machine learning and deep learning techniques to enhance the integrity of GNSS systems.
GNSS signals are essential for modern society, powering applications like navigation, timing, and positioning. As the use of these systems grows, so do concerns about their vulnerability to attacks. The researchers’ work offers a promising solution to this problem, providing a robust framework for detecting and classifying GNSS signals in real-world scenarios.
The team’s findings have important implications for industries that rely on GNSS technology, such as aviation, finance, and transportation. By developing more effective detection methods, these industries can better protect their systems from malicious attacks and ensure the reliability of critical infrastructure.
Cite this article: “Enhancing GNSS Integrity with Machine Learning and Deep Learning Techniques”, The Science Archive, 2025.
Machine Learning, Deep Learning, Gnss, Spoofing, Jamming, Detection, Classification, Convolutional Neural Networks, Ensemble Methods, Transfer Learning.
