The GNSS Spoofing Detection Based On Manifold

Spoofing interference of global navigation satellite system(GNSS)refers to the receiver to capture and track spoofing interference signal of GNSS,so as to get the wrong time,place and other navigation information,resulting in potential risks.How to resist spoofing interference is the guarantee of secure communication in satellite navigation system.The purpose of anti spoofing interference is to identify interference attack and restore true navigation signal.This thesis studies the methods of discrimination of spoofing interference from two aspects,manifold learning and information geometry theory.The existing spoofing interference identification work ignores the difference between the satellite propagation channel and the interference propagation channel,and also does not apply the manifold learning to the spoofing interference identification.This dissertation aims at solving the above problems,and the main work of it is as follows:1.This thesis introduces the global satellite navigation and positioning system and the research significance of anti spoofing interference,and analyzes the present situation of GNSS anti spoofing interference at home and abroad,and introduces the existing methods and techniques of anti spoofing interference.In addition,the theory of information geometry is introduced and the development and application of manifold learning are introduced.2.The concept of manifold is introduced,and the information geometry theory and manifold learning are derived.Information geometry is a theoretical method of statistical analysis on manifold.The purpose of manifold learning is to find manifold structure of high-dimensional data,which is a nonlinear feature extraction method.The basic concepts and main algorithms of the above two methods are introduced in detail.3.This thesis proposes an adaptive neighborhood reconstruction incremental manifold learning algorithm.Taking the ISOMAP algorithm as an example,the problem of incremental manifold learning algorithm is researched after the manifold learning based algorithm of constructing neighborhood problems are solved.From the point of pattern recognition,the algorithm is applied to GNSS spoofing interference recognition,and the features of satellite signal and spoofing interference signal are extracted and classified.Simulation results show the effectiveness of the proposed method.4.Based on the theory of information geometry,a minimum Kullback–Leibler divergence(KLD)detector on statistical manifold is proposed.The Lutz channel model in satellite communication is introduced,and the statistical characteristics of spoofing signal and real signaldue to channel difference are analyzed.The simulation results show that the detector has almost the same performance with the generalized likelihood ratio test detector.