利用支持向量机的矢量跟踪通道故障检测算法  

Applying support vector machine to faulty channel detection in vector tracking loop

作　　者：蒋长辉 陈帅[1] 薄煜明[1] 邓贵军 屈新芬[3] JIANG Changhui;CHEN Shuai;BO Yuming;DENG Guijun;QU Xinfen(School of Automation, Nanjing University of Science and Technology, Nanjing 210094 , China;Graduate Student Brigade, Nanjing Artillery Academy, Nanjing 210000 , China;Institute of Electronic Engineering, China Academy of Engineering Physics, Mianyang 621900 , China)

机构地区：[1]南京理工大学自动化学院,江苏南京210094 [2]南京炮兵学院研究生大队,江苏南京210000 [3]中国工程物理研究院电子工程研究所,四川绵阳621900

出　　处：《国防科技大学学报》2017年第5期21-25,共5页Journal of National University of Defense Technology

基　　金：国家自然科学基金委员会和中国工程物理研究院联合基金资助项目(U1330133);中央高校基本科研业务费专项资金资助项目(30916011336);中国博士后科学基金资助项目(2015M580434);中国博士后特别资助项目(2016T90461);江苏省博士后科研资助计划(1501050B)

摘　　要：为了克服卫星导航矢量跟踪接收机中故障通道对其他正常通道的影响,提高矢量跟踪接收机的鲁棒性,提出一种利用支持向量机的通道故障检测算法,利用矢量跟踪接收机的导航滤波器的新息序列作为支持向量机的输入。支持向量机的输出为通道的状态,根据支持向量机的输出通道的状态来判断该通道是否纳入导航滤波器,用来跟踪滤波器状态值,这样能够有效地避免故障通道对导航结果的污染。仿真实验结果表明:该方法能准确地检测出有故障的通道,提高矢量跟踪环路的鲁棒性。In order to overcome the faulty channels5influence on the receiver and improve the robustness of the vector tracking loops,a faulty channel detection algorithm,which employs the innovative sequences of the navigation filter as the input vector of a support vector machine,was proposed.A trained support vector machine was used to classify the channel as faulty or not faulty.If the channel is classified as faulty,it will not be included in the navigation filter to update the navigation results.In this way,the algorithm can eliminate the bad influence of the faulty channels.The simulation results show that this method can accurately detect faulty channels,and improve the performance and robustness of the vector tracking loops.

关 键 词：卫星导航 矢量跟踪 支持向量机 通道检测 

分 类 号：TN95[电子电信—信号与信息处理]