水下地磁导航实时量测野值的离线模式辨识  

Outlier Identification in Offline Mode for Dynamic Measurements in Underwater Geomagnetism Navigation

作　　者：林沂[1] 晏磊[1] 童庆禧[1,2] 

机构地区：[1]空间信息集成与3S工程应用北京市重点实验室,北京大学遥感与地理信息系统研究所,北京100871 [2]中国科学院遥感应用研究所,北京100101

出　　处：《武汉理工大学学报》2008年第9期112-115,共4页Journal of Wuhan University of Technology

基　　金：国家863课题(2007AA09Z201)

摘　　要：水下地磁导航因其无源性、广域性等优点成为目前导航领域研究的前沿课题,但地磁场的不稳定性使得相关匹配定位过程易受实测粗差的影响而发散。鉴于水下载体运行较慢和通常的海洋磁测仪器可快速测量的特点,结合地磁量测序列为均值非平稳过程的数学归纳,提出水下地磁导航实时量测野值的离线模式辨识算法。该算法包含匹配序列末点前分段容错检测和匹配序列末点一步预测判定2个环节。仿真试验结果表明,该方法可有效增强水下地磁导航相关匹配定位中的发散抑制能力。Underwater geomagnetism navigation now is advanced as a leading subject in navigation research field, but outliers in dynamic measurements may cause positioning results＇ divergence in correlation matching process. Based on the specialties of underwater vehicles＇ slow sailing and marine magnetometers＇ quick sampling, and the reduction of real-surveyed total intensity sequence in non-stationary stochastic process, this paper proposes the combination outlier identification algorithm in offline mode （COIO）. COIO comprises segmentation fault-tolerant outlier identification method （SFTOI） for the sections before matching sequence＇s end point, with one-step prediction judgment （OSPJ） implemented after the end point. Simulation results show that the divergence problem of correlation matching can be avoided effectively with the suggested algorithm.

关 键 词：水下地磁导航 野值检测 离线 容错 

分 类 号：P229[天文地球—大地测量学与测量工程]