Abstract
- GPS spoofing and jamming pose significant threats to maritime navigation, particularly in remote regions where alternative verification methods are unavailable. This paper presents a machine learning-assisted localization system that leverages deterministic satellites and Doppler shift analysis to provide an independent and tamper-resistant positioning method. By analyzing Doppler frequency variations from known satellite trajectories, the system estimates vessel locations and detects GPS anomalies. A machine learning model is trained on satellite observation data to refine location predictions, while a hybridization algorithm optimally combines Doppler-based and ML-generated estimates. Experimental results demonstrate that this approach significantly improves location accuracy compared to standalone Doppler analysis. The proposed method provides a robust GPS-independent localization solution for maritime navigation, enhancing security against GPS spoofing and jamming attacks.