A Deep-Learning-Based GPS Signal Spoofing Detection Method for Small UAVs

The navigation of small unmanned aerial vehicles (UAVs) mainly depends on global positioning systems (GPSs). However, GPSs are vulnerable to attack by spoofing, which causes the UAVs lose their ability. To address this issue, we propose a deep learning method detect spoofing GPS signals received UAVs. Firstly, describe signal dataset acquisition and preprocessing methods; these include hardware system UAV jammer used in experiment, time weather conditions data collection, use Spearman correlation coefficients for preprocessing, SVM-SMOTE solve imbalance. Next, introduce PCA-CNN-LSTM model. We principal component analysis (PCA) model extract feature information related from dataset. convolutional neural network (CNN) was local features dataset, long short-term memory (LSTM) as posterior module CNN further processing modeling. minimize randomness chance simulation experiments, 10-fold cross-validation train evaluate computational performance our machine conducted series experiments numerical environment evaluated proposed against most advanced traditional models. results show that achieved highest accuracy (0.9949). This paper provides theoretical basis technical support detection small-UAV signals.