基于降阶卡尔曼滤波算法在组合导航上的应用  

The application of reduced order Kalman filter algorithm in integrated navigation

作　　者：胡彬[1] 杨新民[1] 王胜红 

机构地区：[1]南京理工大学,江苏南京210094 [2]淮海工业集团,山西长治046000

出　　处：《电子设计工程》2017年第15期98-101,共4页Electronic Design Engineering

摘　　要：本文在简单介绍了卡尔曼滤波基本原理基础上,基于分段线性系统理论以及SOM方法对系统状态可观性和可观度的分析,采用降阶卡尔曼滤波算法,合理剔除不可观或观测度较低的状态因子,对高阶系统进行降阶设计。采用Visual Studio 2010语言开发环境,对常规卡尔曼滤波和降阶卡尔曼滤波进行了算法仿真,通过对前后误差波形的分析比较,结果表明降阶后的滤波系统继承了传统卡尔曼滤波高精度的优点,同时,算法阶数的降低,显著减少了导航计算机的计算负担,实时性得到了显著增强,更易于数字化实现,具有重要实际意义。This paper briefly introduce the basic principle of Kalman filter and analyze the observability and the degree of observability of system based on the piecewise linear system theory and the SOM method. Using the reduced order Kalman filter algorithm to reasonably eliminate the state factors which are not observable or low observable and make a reduced-order design for the high-end system.Using the reduced order Kalman filter algorithm to reasonably eliminate the state factors which are not observable or low observable and make a reduced-order design for the high-end system.At the same time, the reduction of the degree, which has the important practical significance, significantly reduces the computational burden of navigation computer and enhances the real-time so that the algorithm can be more easily realized digitally.

关 键 词：惯性导航 组合导航 卡尔曼滤波 数字仿真 

分 类 号：TN961[电子电信—信号与信息处理]