利用CNN-LSTM融合模型实现GNSS诱导式欺骗干扰检测

打开文本图片集

关键词：欺骗干扰检测；卫星导航；诱导式欺骗；卷积神经网络；长短期记忆网络；残差网络中图分类号：TN967.1-34；TN911.7 文献标识码：A 文章编号：1004-373X（06）07-006-05

GNSS induced spoofing jamming detection using CNN-LSTM fusion model

SUNMingzhe¹，WANG Zhenling'，HAO Fang1，² （1.The54th Research InstituteofChina Electronics Technology GroupCorporation，Shijiazhuang O5o081，China; 2.School of Instrument Scienceand Engineering，Southeast University，Nanjing 21Oo96，China）

Abstract：Satelitenavigationreceivershavelimitedcapabilities incountering inducedspofingaacks，andthetraditional detectionmethodsusedarefacedwithchallngessuchasreal-timeprocessingdiffcultiesandpooradaptabilitytopreset discrimination thresholds.Inviewofthis，thepaperproposesafusionneural networkdetectionmethodbasedonCNN-LSTM. Firstlythecoelationpeakaliasingharactersticsduringthspooingpuloffpaseasaad.Ten，teResNet-18susd asthebackboneoftheconvolutionalneuralnetwork（NN）toextractspatialfeaturesinthecodephasedomainandDopplerdomain， andthelong short-term memory （LSTM）network was employed to track the temporal dependenciesacrossconsecutive frames，so astodetecttheinducingbehaviorofdeceptivesignals.Tosimulatetheinducedspoofingprocess，acorrelationambiguity function（CAF）sequencedataset wasconstructedtoverifythedetectionperformanceof thefusionmodel.Experimentsshowthat thedetectionaccuracyrateof theproposed methodforinducedspoofingatacksexceeds98%，whichisimprovedby2%than thatof the traditionalsinglemodels.Moreover，boththedetectiondurationandmodelcomplexitycanmettherequirementsof civilianreceivers.Tosumup，itisan effective method inthe fieldof anti-spoofing applicationofsatellite navigation.

Keywords：spoofing jamming detection; GNSS; induced spoofing; CNN; LSTM network; residual network

0 引言

随着全球导航卫星系统（GNSS）的发展，卫星导航技术已在军事和民用领域得到了广泛应用。（剩余7691字）