基于奇异值分解的改进机载单站无源定位算法  

An Improved Algorithm of Single Airborne Observer Passive Location Based on SVD

作　　者：周峰[1] 焦淑红[1] 孔挺[2] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001 [2]海军飞行学院教研部,辽宁葫芦岛125001

出　　处：《探测与控制学报》2011年第3期19-23,共5页Journal of Detection & Control

基　　金：973国家安全重大基础研究基金项目资助(61393010101-1)

摘　　要：针对机载单站无源定位系统中的滤波算法存在滤波稳定性差、收敛速度慢、定位精度差等问题,提出一种基于奇异值分解的平方根sigma点卡尔曼滤波算法(Square Root Sigma Point Kalman filter based on Sin-gular Value Decomposition,SVD-SRSPKF)。新算法利用奇异值分解代替Cholesky分解或更新,并使用误差协方差的平方根替代协方差进行滤波,保证滤波算法的数值稳定性。仿真结果表明:SVD-SRSPKF算法比其他同类算法具有更高的收敛速度、定位精度和数值稳定性。In order to solve the problems of poor filtering stability,slow convergence speed and low locating accuracy of the filtering algorithm in the single airborne observer passive location,a square root sigma point Kalman filtering algorithm based on singular value decomposition（SVD-SRSPKF） was proposed in this paper.The Cholesky decomposition or update was replaced by singular value decomposition,and the square root of covariance was used to d the filter,which ensured filtering algorithms numerical stability in the new algorithm.The simulation results showed that the SVD-SRSPKF algorithm had higher convergence speed,location accuracy and numerical stability than any other similar algorithm.

关 键 词：机载单站无源定位 奇异值分解 平方根滤波 数值稳定性 

分 类 号：TN958.97[电子电信—信号与信息处理]