Performance Testing and Analysis of a New GNSS Spoofing Detection Method in Different Spoofing Scenarios

This paper addresses the vulnerability of Global Navigation Satellite Systems (GNSS) to spoofing signal attacks by proposing a spoofing detection method based on multi-parameter features and an optimized random forest algorithm. Traditional methods, which rely on a single parameter, struggle to adapt to diverse spoofing strategies and suffer from limitations in reliability and robustness. To overcome these shortcomings, this study extracts multi-dimensional parameters from observational data. By improving the RF algorithm and introducing a weighted voting mechanism to optimize the classification decision process, a high-precision classification model is constructed. Experiments were conducted using the Texas Spoofing Test Battery (TEXBAT) dataset and a self-built spoofing platform to evaluate the performance of the proposed method under various spoofing scenarios. The results demonstrate that the optimized Random Forest achieves a detection probability exceeding 97% in both single and mixed scenarios. Furthermore, the method exhibits stable performance in dynamic generated spoofing scenarios, with a detection accuracy rate surpassing 99%. The study proves that the combination of multi-parameter joint analysis and optimized Random Forest effectively enhances the generalization capability and anti-interference performance of GNSS spoofing detection, providing a new approach for navigation security in complex environments.