|本期目录/Table of Contents|

[1]周磊,董乃铭,洪振杰.UKF 算法与SVDKF 算法性能的比较[J].温州职业技术学院学报,2013,01:81-83,86.
 Zhou Lei,Dong Naiming,Hong Zhenjie.Comparison of Algorithm Performances of UKF and SVDKF[J].Journal of Wenzhou Vocational and Technical College,2013,01:81-83,86.
点击复制

UKF 算法与SVDKF 算法性能的比较(PDF)

《温州职业技术学院学报》[ISSN:1006-6977/CN:61-1281/TN]

期数:
2013年01期
页码:
81-83,86
栏目:
应用技术
出版日期:
2013-03-15

文章信息/Info

Title:
Comparison of Algorithm Performances of UKF and SVDKF
作者:
周磊董乃铭洪振杰
温州大学 数学与信息科学学院,浙江 温州 325035
Author(s):
Zhou Lei Dong Naiming Hong Zhenjie
School of Mathematical and Information Science, Wenzhou University, Wenzhou, 325035, China
关键词:
UKF 算法SVDKF 算法滤波Cholesky 分解SVD
Keywords:
UKF SVDKF Filter Cholesky decomposition SVD
分类号:
O235
DOI:
-
文献标识码:
A
摘要:
在模式识别领域,基于Unscented 的卡尔曼滤波算法(UKF)广受关注,但在求解过程中经常会遇到病态问题,从而影响算法的性能。基于奇异值分解(SVD)的卡尔曼滤波算法(SVDKF)以SVD 代替Cholesky 分解协方差矩阵产生sigma 样本点,可以提高协方差矩阵的数值稳定性。通过对两种算法性能进行仿真比较发现,SVDKF 算法优于UKF 算法,具有良好的鲁棒性,能有效改善滤波性能,提高算法的精度。
Abstract:
In the field of pattern recognition, Kalman filter algorithm (UKF) based on Unscented has been widely used but often encounters the ill-conditioned problems in solving problems, which affects the performance of the algorithm. A Kalman filter algorithm (SVDKF) based on SVD (singular value decomposition) generates sigma samples by using singular value decomposition instead of Cholesky decomposition, and it improves the numerical stability of the covariance matrix. Through simulation experiment on their performances, it is found that SVDKF with higher robustness is better than UKF, and it can effectively improve the performance and accuracy.

参考文献/References

[1] Uhlmann J K.Algorithm for multiple target tracking[J].American Science,1992,80(2):128-141.
[2] Chen Z,Haykin S.Bayesian filtering:From Kalman filters to particle filters,and beyond[R].Adaptive Systems Lab:McMaster University,2003.
[3] Youmin Zhang,Guanzhong Dai,Hongcai Zhang.A SVD-based extended Kalman filter and applications to aircraft flight state and parameter estimation[C]//Proceeding of the American Control Conference.Piscataway:IEEE Press,1994:1809-1813.
[4] Julier S J,Uhlmann J K,Durrant-Whyte H F.A new method for the nonlinear transformation of means and covariances[J]. IEEE Transactions on Automatic Control,2000,45(3):477-482.
[5] Liu Bin,Ma Xiao-chuan,Hou Chao-huan.A particle filter using SVD based sampling Kalman filter to obtain the proposal distribution[C]//International Conference on Cybernetics and Intelligent Systems,2008:581-584.
[6] John L Crassidis,F Landis Markley.Unscented Filtering for Spacecraft Attitude Estimation[J].Journal of Gu ldance and Dynamics,2003,26(4):536-542.
[7] Van der Merwe R,Wan E A.The square-root unscented Kalman filter for state and parameter estimation[C]//Proceedings of the International Conference on Acoustics,Speech,and Signal Processing(ICASSP),2001:3461-3464.
[8] Xiaoqin Zhang,Weiming Hu.SVD based Kalman particle filter for robust visual tracking[C]//19th International Conference on Pattern Recognition(ICPR),2008:1-4.
[9] XIE Qiang,TANG He-sheng,DI Yuan.SVD-unscented kalman filter for nonlinear structural system identification[J]. Chinese Journal of Applied Mechanics,2008,25(1):57-61.

备注/Memo

备注/Memo:
-
更新日期/Last Update: 2013-03-20