Small Delay GNSS Forwarding Spoofing Detection in a Multipath Environment Based on Convolutional Neural Network

Global navigation satellite systems (GNSSs) are popular for navigation and timing applications due to their high accuracy, low cost, and global availability. However, the open signal structure and extremely low signal power make them susceptible to various intentional and unintentional interferences. Spoofing can manipulate the navigation information of the victims, which is stealthier and more harmful. In practical environments, the received signals often contain multipath signals in addition to the direct path, which are caused by reflection or scattering from obstacles. Among them, both small delay multipath signals and small delay forwarding spoofing signals can distort the capture correlation peaks, leading to a high detection false alarm rate of existing spoofing detection methods based on the shape of correlation peaks. In this article, we address this problem by analyzing the positional differences of the correlation peaks of different signals in the capture stage relative to the capture center based on the time delay and power characteristics. Subsequently, using a convolutional neural network (CNN) effectively detects small delay spoofing signals in a multipath environment. When the superimposed power of the multipath signals in the receiving environment is less than that of the genuine signal, the proposed method can effectively distinguish the multipath signal from the spoofing signal. In addition, the spoofing detection accuracy of the CNN on the simulation dataset reaches 98.83%. Finally, the effectiveness of the proposed method is verified by building field experimental scenarios.