舰船组合导航系统与顺序滤波融合算法[J].docVIP

3 舰船组合导航系统的顺序滤波算法 摘 要:一阶Markov过程舰船SINSGPS组合导航关键词: 组合导航集中式融合顺序滤波中图分类号： 文献标识码：A Sequential Filtering Fusion AlgorithmIntegrated Navigation System TANG Tian-hao, WEN Chun-hong, WANG Tian-zhen, GE Quan-bo (Shanghai Maritime University, Shanghai, 200135) Abstract: Centralized fusion and distributed fusion for vessel integrated navigation have either higher filtering accuracy or better computation performance and fault-tolerant, but not both. And impacted by marine environment, it is too ideal to modeling vessel process noise with white noise. To against above problems, the paper introduces sequential filtering fusion algorithm for integrated navigation system. The new algorithm models vessel process noise with one order Markov process, and changes state equations into basic equations of standard Kalman filter, then checks the efficiency of each sub navigation system and fusions all efficient data step by step to positioning vessel. Compared with traditional sequential filtering fusion algorithms, the new algorithm reserves advantages of the same filtering accuracy with centralized fusion algorithm and better computation performance, and adds advantages of more practical and better fault-tolerant. Theoretical analysis and vessel GPS / SINS integrated navigation simulation results indicate the efficiency and superiority of the new algorithm. Key words: integrated navigation; centralized fusion; distributed fusion; fault-tolerant; sequential filtering fusion; state dimension augmentation 引言 全球化进程的加快促进了船舶运输业的迅猛发展，从而使得船舶对导航定位设备的自动化和精度提出了更高的需求。近年来，信息融合技术的进步极大地推动了舰船组合导航技术研究和应用的发展，取得了诸多研究成果，其中部分已在实际工程中得到了有效应用[1-]。 舰船组合导航系统中常用的融合方法包括集中式和分布式两种。集中式具有高的精度，但要求融合滤波器具有强的中心处理能。该算法的基本思想为：1)用一阶Markov过程建状态方程为符合标准Kalman滤波的基本方程；2)顺序滤波实现。状态方程为 (1) 其中表示离散时间标记；为目标的状态矢量，为状态转移矩阵船舶行受海洋环境的影响，舰船组合导航系统中的运动过程噪声一阶Markov过程建模更实际即 () 上式中，是方差为的零均值高斯白噪声 舰船上的各导航子系统(如GPS、SINS、TNS、CNS和SAR等)经过时空配准后的测量方程写为 , () 上式中，是导航子系统的数目，测量噪声是相互独立的零均值高斯白噪声过程，其方差为噪声方差阵为 () 1.2 舰船组合导航系统模型的标准化 易知，1.1节