Adaptive GNSS Anti-Spoofing Algorithm Design And Implementation For Self-Driving Cars

The Global Navigation Satellite System(GNSS)has become an essential component of autonomous driving technology as it develops,and many related technologies depend on GNSS positioning.However,in actual contexts,GNSS is vulnerable to outside interference and deliberate spoofing,and incorrect positioning may have life-threatening consequences for self-driving vehicles.Numerous academics have addressed this topic,primarily using signal processing or the incorporation of additional sensor data sources for cross-validation.The former may require the introduction of additional hardware,which adds cost to the industry.The latter faces the problem of how many additional data sources must be introduced for verification.Fewer data sources do not guarantee the detection of all spoofing,and more data sources impose a computational burden on self-driving vehicles.In order to solve the above problems,this paper introduces factor graphs into GNSS spoof detection for the first time to pursue a balance between detection effectiveness and computational consumption.With the plug-and-play capability of factor graphs,it can adaptively fuse sensor data for detection according to the needs of different scenarios.This initiative reduces the consumption of computational power on board while ensuring the detection effect.The main contributions of this paper are as follows:· A scenario testing and context awareness approach based on digital twins:In this paper,we propose a digital twin-based approach for scenario testing and context-awareness.The method is based on the obtained scenario data and simulates the test in a virtual model by digital twin,and the probability of known risk is obtained after analysis and calculation.After that,a risk graph model based on Bayesian networks is suggested,and the probability that unknown dangers would occur is examined.The probability of occurrence of the unknown risk can be used to evaluate the sensor state,and the digital twin can perceive the context in which the physical object is located based on the sensor state,thus providing analysis for later factor graph-based adaptive GNSS spoofing detection and defense.· Framework for GNSS spoofing detection and defense based on factor graphs:In this paper,we propose an adaptive GNSS spoofing detection and defense framework based on factor graphs.Due to the plug-and-play advantage of factor graphs,the detection framework can well switch the strength of detection required in the current state based on the perceived context,i.e.,the number of sensors incorpo-rated into the computation,thus saving the arithmetic power used for detection computation when the risk is small and incorporating more data for ensuring the detection rate of spoofing attacks when the risk is large.In addition,the estimation of factor graphs will be more accurate compared to detection schemes based on extended Kalman filtering,which means that more advanced attacks can be detected.· GNSS spoofing defense methods:This paper presents a means of defense after GNSS spoofing is detected.The current stage of work focuses on the detection of spoofing,and the detection of spoofing can only be handed over to the system or manually as a warning.The defense proposed in this paper,on the other hand,can use other reliable sensor data on the fusion after GNSS spoofing is detected to estimate a vehicle position with the same high accuracy as the localization output of the self-driving vehicle.