Neural Network Based C/N0 Abnormity Detection Method for GPS Anti-spoofing

Recent years have witnessed the emergence of a new type of threat to the civilian GPS, namely spoofing. Spoofing is a concealed sabotage that coerces victim receiver into providing an erratic position result without any warning. Usually, spoofing is completed by transmitting fake GPS signal at a slightly higher power level to the receiver antenna from nearby. There have been many techniques proposed for detection and mitigation of the spoofing signal. However, these methods often focus on one aspect of characteristics, neglecting the consistency within GPS signals. In this paper, we explore the relationship between the mean carrier to noise ratio value and the angular position of corresponding satellite. This relationship is quite stable for fixed receiver, and forms a spatial pattern. The uniqueness and complexity of this pattern qualifies itself as an adequate fingerprint of the authentic signal, therefore it has the potential for discriminating spoofing signal. We propose a method based on artificial neural network, which could detect the abnormity in this spatial pattern. Some points of implementation and experiment results are presented. At the end, the advantage and potential of this cross-level information fusion framework is discussed in depth.