水下无源组合导航系统智能容错方法设计  

Design of intelligent fault-tolerant to passive underwater integrated navigation system

作　　者：李佩娟[1] 徐晓苏[1] 王立辉[1] 李瑶[1] 王立辉 

机构地区：[1]微惯性仪表与先进导航技术教育部重点实验室,南京210096 [2]东南大学仪器科学与工程学院,南京210096

出　　处：《中国惯性技术学报》2013年第2期221-225,共5页Journal of Chinese Inertial Technology

基　　金：国家自然科学基金项目(60874092,50575042);江苏省博士后科研资助(1202007C);东南大学基本科研业务费资助(3222003061);东南大学微惯性仪表与先进导航技术教育部重点实验室(B类)开放基金资助项目(3222002301)

摘　　要：为提高自主水下航行器的导航精度,比较目前AUV常用的水下导航方式,将捷联式惯性导航系统与地球物理导航系统相结合构成水下无源组合导航系统。采用容错联邦卡尔曼滤波对各子系统信息进行故障诊断、系统重构和融合。针对传统的2检验法不能确定故障具体原因,而仅能判断量测信息是否有效的缺陷,提出利用神经网络辅助2检验法进行故障诊断。通过对水下组合导航系统算法进行仿真分析,结果表明该算法能够快速、准确地判断系统故障源,通过故障隔离和系统重构,使系统在故障情况下依然保持正常工作。To improve the navigation accuracy of autonomous underwater vehicle, an integrated navigation system was proposed, which was composed of strapdown inertial navigation system （SINS）, terrain-aided navigation （TAN） system and gravity-aided navigation （GAN） system. A fault-tolerant federated Kalman filter was used to fuse the various navigation sensors, detect the system fault and reconstruct the navigation system. In view that traditional cai-square hypothesis testing processes fault detection cannot determine the specific cause of the fault, and could only determine the validity of measure information, a new fault detection algorithm based on neural network was adopted. The application on underwater integrated navigation system demonstrates that the algorithm can rapidly and accurately detecting and identify the faults in the system. Therefore the effective fault isolating can be performed to realize the fault tolerance navigation.

关 键 词：捷联惯性导航系统 地球物理导航 联邦卡尔曼滤波 故障诊断 神经网络 

分 类 号：U666.1[交通运输工程—船舶及航道工程]