Protective system for drones against GPS spoofing using image and signal analysis

Drones play a critical role in air navigation, serving various safety-of-life and positioning, navigation, and timing applications. However, the increasing use of Drones has led to a rise in GPS spoofing attacks. This work aims to protect Drone GPS systems by enabling the detection of spoofing attacks. The proposed solution involves utilizing a pretrained ResNet36 model for image analysis, comparing images captured by drones with satellite images of the same locations. Simultaneously, we explored RF signal processing using a diverse dataset containing both clean and distorted signals. By integrating PCA data with a LSTM, our solution achieved an impressive 100% accuracy in classifying spoofed signals.