Research On GNSS Spoofing Detection Algorithm And Experimental Verification Methods

This paper mainly studies the detection algorithm of GNSS spoofing signal and the inertial signal/satellite spoofing signal simulation.With the development of satellite navigation technology,especially China's Beidou satellite navigation system coverage and application areas are expanding,intentional or unintentional spoofing will affect the application security of GNSS.In particular,intentional spoofing will directly lead to the "wrong" location information of the positioning terminal and the "wrong" time information of the timing terminal,disrupting the normal production and life order.Based on the actual requirements,this paper firstly studies how to detect spoofing signals under the existing conditions(i.e.,without changing the terminal hardware design).Based on the different signal transmission paths,the pseudorange models of real signal and spoofing signal are established.Furthermore,the pseudorange difference models of real signal scene and spoofing signal scene are established.Firstly,a spoofing signal detection algorithm based on dual receivers is proposed in this paper.By difference calculation and iterative solution,the distance between the position results of two receivers is compared with the real baseline distance,so as to realize the discrimination of the signal authenticity.For the single-receiver double-difference detection algorithm,the pseudorange double-difference(time difference and inter-satellite difference)model is linearized by Taylor approximation expansion,and the theoretical expression of the position solving algorithm based on pseudorange time difference and inter-satellite difference is obtained.Comparing it with the theoretical expression of the least squares method based on pseudorange,we deduce and analyze the difference between these two algorithms in the condition of different signal scenes,which provides support for confirming the authenticity of signals.For the single receiver single-difference detection algorithm,the pseudorange single difference(time difference)model is linearized by Taylor approximation expansion.The difference between the position solving algorithm based on pseudorange time difference and the least squares method based on pseudorange is analyzed in view of the real signal scene and the spoofing signal scene respectively.The difference of these two algorithms is analyzed theoretically.For three different detection schemes,the criterion of signal authenticity is given.At the same time,based on this criterion,by carrying out multiple sets of real experiments including the college building roof experiment,the corridor experiment,the school playground experiment,the Beidou open laboratory experiment,the Chengdu experiment and the TEXBAT data sets experiment,we study and verify the performance of the algorithm in the condition of the real signal,repeated spoofing signal,generated spoofing signal,static,dynamic,full time spoofing and so on.The simulation results verify the feasibility and effectiveness of the algorithm.In addition,the simulation analysis of the spoofing signal detection method based on CN0 hopping is carried out,which further illustrates the advantages of the proposed algorithm.In the condition of full time spoofing,the influence of the initial value setting on the detection performance of the algorithm is discussed when the receiver is in static and moving states and facing repeated spoofing and generated spoofing respectively.Based on the idea of inertial assist,the signal detection algorithm based on pseudorange difference,the signal detection algorithm based on distance difference and the signal detection algorithm based on single-satellite tight integration(position difference,the variation of the receiver clock offset/clock drift have a certain regularity)are proposed.The signal detection algorithm based on pseudorange difference mainly discriminates the real and the spoofing signals by comparing the pseudorange measured by the receiver with the reference pseudorange constructed by the inertial navigation solution and the satellite position.The signal detection algorithm based on distance difference compares the differences and similarities of the distance between the satellite navigation solution and the inertial navigation solution in different signal scenarios at the same time and at different times.The signal detection algorithm based on the single-satellite tight integration realizes the signal detection through the tight integration of each satellite and the inertial navigation system and comparing the integration results(including position and the equivalent distance deviation of the receiver clock offset/clock drift).The four algorithms analyze the different performance of the real signal and the spoofing signal.The advantage of these four algorithms is that they can not only detect the whole-satellite spoofing,but also effectively deal with the detection of the partial satellite spoofing.By eliminating the satellite signals affected by spoofing and adding the remaining unaffected satellite signals into the integrated navigation solution,the error accumulation of inertial navigation can be slowed down and the positioning accuracy of the integrated navigation system can be guaranteed.In the simulation of satellite spoofing scene,the relationship between velocity and position in the navigation coordinate system is analyzed.By taking the altitude deviation of 500 m and 1000 m as the target,the change of pseudorange and pseudorange rate by the spoofing signal relative to the real signal is deduced.Based on two sets of measured data,the signal data satisfying the spoofing purpose was generated from static and dynamic aspects,which provides the data support for the simulation anslysis of the spoofing detection algorithms.In the simulation of inertial navigation IMU signal,in order to match the carrier's kinematics and dynamics characteristics,this paper studies the simulation problem with high precision and high renewal rate of the gyro angle increment and the accelerometer specific force integral increment based on the actual data.In the inertial coordinate system,IMU signal generation algorithms based on quaternion and dual quaternion are proposed.For the signal generation algorithm based on quaternion,the attitude quaternion and its derivative,position and speed are taken as the endpoint value and endpoint derivative value respectively in the new time interval,and the new cubic spline polynomial can effectively reduce the influence of the rounding error when solving the integral.The first and second order Taylor approximation methods are proposed to deal with the fraction terms in the analytic expressions of the gyro angular rate and the accelerometer specific force.The equivalent scale factor errors of the first and second order approximations are analyzed,and the accuracy of the generated 100 Hz,200Hz and 400 Hz data is analyzed.For the signal generation algorithm based on dual quaternion,the dual part of the dual quaternion is represented by velocity translation vector.The analytical expression of the gyro angular rate and the accelerometer's specific force is obtained by constrained interpolation.The average approximation method is proposed to deal with the fraction term in the expression.The algorithm error caused by approximation and its solution are given.By the simulation analysis of actual flight data,the accuracy of the data generated by the quaternion-based algorithm,the dual quaternion-based algorithm and the inverse SINS algorithm is compared.The simulation results show that the accuracy of the generated signal can fully meet the requirement of offline simulation.