Detection Of Navigation Signals Spoofing Based On Neural Networks

The navigation and location services of the global satellite navigation system have brought people an unprecedented good experience,but firstly because the navigation signals convey the ground to go through a long process,Secondly,the openness of the navigation signal and its lack of protective measures make the navigation signal very vulnerable and easy to be deception and interfered by malicious users.This type of interference is different from the suppressive interference which is easy to be detected,it is not only highly concealed,but also extremely harmful,and it is difficult to be detected by ordinary instruments and algorithms.Aiming at the above problems,a detection method of navigation signals spoofing based on BP neural network supervised machine learning to classify deceptive navigation signals and real navigation signals is proposed.Based on the original vulnerability analysis,it fills the blank of the vulnerability analysis of the receiver at the signal processing level.This paper analyzes the principle of deception interference and does not divide deception interference into forward interference and production interference according to the generation method.Analyze the three ways to achieve deception interference at this stage,including the use of commercial signal simulators,the use of "FPGA+DSP" architecture to generate analog signals,and the use of software radio for forwarding interference experiments.The experiment is carried out under the UBUNTU system,using navigation messages and three-dimensional coordinates to generate simulation data,which is transmitted through the HACKRF radio frequency module,and then affects the target receiver.The highly nonlinear characteristic of BP neural network makes it possible to approximate any function,so it has a good effect on the processing of nonlinear data and has unique advantages in signal classification.By using field data to generate deceptive interference signals,the use of complex methods to generate deceptive interference is avoided.Using software-defined radio collects the digital intermediate frequency signal of the satellite navigation system at different locations as the input data of the software receiver program.After the software receiver is processed and solved,extracting the observations such as satellite number,pseudorange,carrier phase,doppler frequency shift,receiver clock frequency drift,receiver clock error,and signal-to-noise ratio as features from the output data.And the data set composed of these features is used as the input of the BP neural network.Finally,the trained neural network model is used for classification test,and the deception interference detection is completed.It can be seen from the simulation results that the classification effect of this method reaches 84.07%,indicating that it has high detection performance and can be further studied.