Research On GNSS Anti-spoofing Scheme Based On Particle Filter

Global Navigation Satellite System (GNSS) is widely applied in aerialnavigation, atmospheric monitoring, geology exploration, etc, provid ing thelocation, navigation and timing service. However, because of the vulnerability ofGNSS, its potential risk attracts the researchers’ attention. Spoofer repeats orgenerates the spoofing GNSS signa ls, which are similar to the authentic signals. Asthe spoofing signa ls contain the fake ephemeris and pseudo-ranges, they canmislead the calculation after they were tracked. Based on the distributed particlefilter, we proposed an improved Receiver Autonomous Integrity Monitoring (RAIM)scheme, which can improve the accuracy of positioning and timing, as well aseliminating the influence of spoofing.Firstly, we introduce the theory of least square calculation for positioning andvelocity. By the measured pseudo-ranges, the measured variations of pseudo-ranges,the positions of satellites and the velocity of satellites, the receiver can set up anon-linear equation set related with the receiver’ position and velocity. Classicalleast square calculation utilizes the Newton iteration to linearize the non-linearequation set, and estimates receiver’s position and velocity by least square.Additionally, we also introduce the ionosphere delay, troposphere delay, receiver’snoise and some other error terms, and analyze their influences for the calculationresults.To improve the accuracy of positioning and timing, we use the partic le filterfor calculation, and propose distributed partic le filter. Partic le filter is anapproximately optima l Bayesian estimator based on Monte Carlo simulations. Bythe sequentia l importance sampling, partic le filter can accurately estimate the statevector even in the non-linear and non-Gaussian space, Hence the GNSS calculationbased on the partic le filter can provide better accuracy than least square calculation.Furthermore, we propose distributed particle filter, which contains one mainsub-filter is for positioning, and the other one assistant sub-filter is for velocityestimating. Particles for each sub-filter are both four-dimensiona l vectors, whichcan cover the state space more adequately compared to one singleeight-dimens ional vector with the same number of the particles. Therefore, theproposed filter can further improve the positioning accuracy. Considering the vulnerability of GNSS, we proposed an improved ReceiverAutonomous Integrity Monitoring (RAIM) scheme to resisting the spoofing attack.By theoretical analysis, we find that the additional ranges are the key factors ofpositioning misleading, and we classify the spoofing attack by the number ofspoofing satellites and the variation rate of additiona l ranges. We firstly assess thethreat of one and multip le spoofing satellite attack in the scenario of jump ingadditional ranges and slowly changing additional ranges, and present the possibleattack modes in the perspective of spoofer. Then an improved RAIM is proposed forspoofing detection and identification. The detection is based on the magnitude ofmeasurement vector, and the identification is base the linear correlation ofmeasurement vector. By virtue of the improved RAIM, receiver can correct thepositioning under the spoofing attacking.