INS/GNSS Integrated Navigation System Spoofing Detection Techniques

While the GNSS spoofing threat is emerging,the security and reliability of navigation system face new challenges.Since GNSS spoofing signals are almost identical to the authentic ones,such concealment makes it difficult to guarantee reliable spoofing detection,simply relying on GNSS receivers.GNSS spoofing attacks can adversely affect the entire navigation system and related applications by controlling the position/time solution of the GNSS receiver:in the military field,erroneous navigation and positioning information will greatly weaken the operational effectiveness of various weapons;in the civilian sector,it poses a serious threat to safety-critical applications such as air navigation.With INS/GNSS integrated navigation systems gradually becoming the basic configuration for vehicle navigation,the combination of INS/GNSS,atomic clock and dual antenna is an effective way to improve GNSS spoofing detection ability.From the perspective of INS/GNSS integrated navigation,this dissertation studies and proposes three spoofing detection algorithms based on filter innovation,chip-scale atomic clock(CSAC)and dual antenna configuration,from the position dimension,time dimension and signal-space-geometry dimension.Combined with theoretical analysis and field experiments,the effectiveness of these spoofing detection methods is verified.The main contents of this dissertation are as follows.1.The classification of GNSS spoofing attacks and characteristics of different spoofing modes are summarized and compared.Considering the spoofer’s signal processing,signal transmission,data bit prediction and controllable delay,the mathematical correlation model of spoofed and authentic pseudorange measurements is established,which clarifies the impacts of typical spoofing attack modes on GNSS measurements and solutions.The effects of spoofing attacks on Kalman filter covariance,innovation and inertial sensor bias estimation are analyzed.Theoretical analysis and simulations show that spoofing attacks have no direct effects on filter gain calculation loop,so the error covariance matrix cannot reliably reflect the filter’s performance;spoofing profiles have significant influence on the expectation of innovations in the initial stage,but the influence is gradually weakened due to the filter’s error tracking effects,which gives spoofing detection a certain time window;real-time monitoring of inertial sensor bias estimation and backtracking technique can be used as effective means for detecting spoofing attacks and suppressing the introduced adverse effects.2.In the position dimension,there should be a difference between the spoofed GNSS position and INS dead-reckoning position.For the existing innovation averaging algorithm has shortcomings in detecting slowly-drifting spoofing attacks,an innovation-based measurement averaging detection algorithm is proposed.In the given time window,multiple measurements at different epochs are transferred to the same time for measurement update.The tracking of the spoofing profile is suppressed without losing measurement information.The concept of virtual innovation is introduced,which proves the advantage of measurement averaging over innovation averaging in detecting slowly-drifting spoofing errors.For spoofing profiles with small drift magnitude,in the same averaging window,measurement averaging is better than innovation averaging in both spoofing detection probability and detection time.Meanwhile,it is proved that due to spoofing attacks,the receiver measurements change consistently,and thus multiple satellites no longer provide any fault redundancy information,which leads to failure of the widely used aviation integrity monitoring algorithm based on solution separation.3.In the time dimension,the influence of unsynchronized spoofing attacks on GNSS receiver time is inevitable.For spoofing detection based on the receiver’s internal clock suffers performance degradation after a long period of jamming,spoofing detection algorithm based on CSAC’s high-precision timekeeping ability is proposed.Timekeeping error characteristics of CSAC are analyzed theoretically.The disciplining process and timekeeping ability of CSAC are tested and verified.Based on the impact assessment of spoofing attacks on receiver time,a CSAC assisted spoofing detection model is established.Field experiments and theoretical analysis show that,the root mean square error of CSAC clock bias prediction is only a few meters in 20 minutes,which is two orders of magnitude smaller than the receiver’s internal clock;under 200 s jamming,when the spoofing attack distance is 100 m,the theoretical spoofing detection probability based on CSAC is greater than 99.9%,which is better than the receiver’s internal clock(only 60%);under 20 min jamming,when the spoofing attack success rate is 80%,CSAC can limit the effective range of the spoofing attack to 30m;the introduction of CSAC can greatly increase the difficulty of carrying out spoofing attacks,and improve the spoofing detection performance and anti-spoofing capability of the navigation system.4.In the signal-space-geometry dimension,there exists a difference in the angle of arrival between the spoofing signals and the authentic ones.For the hardware complexity or detection performance evaluation deficiency of the present techniques,an INS assisted dual antenna spoofing detection algorithm is proposed.By introducing the INS attitude and position solutions,theoretical values of the double difference observations can be accurately predicted without using the multi-element antenna array.Based on the half-cycle fractional observations of carrier phase measurements,deterministic double difference probability distribution models under authentic and spoofing signals are established.A spoofing detection algorithm with controllable real-time detection probability is designed under constant false alarm rate.Key parameters such as false alarm rate,missed detection rate,and smoothing period of carrier phase/pseudorange double difference spoofing detection are analyzed and compared.Theoretical analysis and experiment results show that when the predicted carrier phase double difference reaches 0.37 cycles,spoofing detection probability greater than 99.9% can be achieved in a single epoch;the pseudorange double difference spoofing detection algorithm can reliably detect spoofing attacks when the smoothing period requirement is met.5.For existing spoofing attack simulation schemes cannot meet the requirements of integrated navigation system spoofing detection algorithm verification,and the cost is high,a dual-channel SDR-based spoofing attack simulation scheme is designed.Satellite signal simulation and high fidelity inertial measurement unit data simulation are combined.A semiphysical platform for integrated navigation system spoofing detection verification is built.Based on the simulation platform,synchronized and unsynchronized spoofing attacks are effectively simulated and verified by real GNSS receivers as attack targets.Finally,the three spoofing detection algorithms proposed in this dissertation are verified and compared under typical spoofing attack scenarios.