基于CS-C-SVM的多参数GNSS欺骗干扰检测  

Multi-parameter GNSS Spoofing Interference Detection Based on CS-C-SVM

作　　者：卢丹[1] 殷亚强 LU Dan;YIN Yaqiang(Tianjin Key Lab for Advanced Signal Processing,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学天津市智能信号与图像处理重点实验室,天津300300

出　　处：《信号处理》2022年第6期1325-1332,共8页Journal of Signal Processing

基　　金：中央高校基本科研业务费项目(3122019050);天津市自然科学基金(19JCQNJC01000)。

摘　　要：欺骗通过传输虚假的全球导航卫星系统(global navigation satellite system,GNSS)信号使接收机解算出错误的位置、速度和时间,具有很大的危害性,因此检测到欺骗干扰是十分必要的。传统的利用单一参数检测欺骗干扰的方法存在一定的局限性,然而欺骗过程会引起一系列参数的变化,本文提出综合利用多个参数的检测方法,即将多个参数作为C-支持向量机(C-support vector machine,C-SVM)的特征输入并构建用于检测欺骗干扰的分类器。考虑到使用传统网格搜索(grid search,GS)优化后的C-SVM算法(GS-C-SVM)容易陷入局部最优继而降低分类器的性能,文中提出利用布谷鸟搜索(cuckoo search,CS)算法优化C-SVM(CS-C-SVM),仿真实验结果表明该方法可以进一步提高分类准确率并降低虚警率。Spoofing makes the receiver calculate the wrong position,speed and time by transmitting false global navigation satellite system(GNSS)signals,which is extremely harmful,so it is very necessary to detect spoofing interference.The traditional method with using a single parameter to detect spoofing interference has certain limitations,but spoofing process makes a series of parameters change.A detection method with comprehensively using multiple parameters was proposed in this paper,which multiple parameters were used as C-support vector machines(C-SVM)feature input and constructed a classifier for detecting spoofing interference.The C-SVM algorithm optimized by the traditional grid search(GS-C-SVM)is easy to fall into the local optimum,decreasing the performance of the classifier.The paper proposed to use the cuckoo search algorithm to optimize the C-SVM(CS-C-SVM).Simulation result shows that this method can further improve the classification accuracy and reduce the false alarm rate.

关 键 词：卫星导航 欺骗干扰检测 多参数 C-支持向量机 布谷鸟搜索 

分 类 号：TN967.1[电子电信—信号与信息处理]