基于GPR模型的自适应平方根容积卡尔曼滤波算法
收稿日期: 2012-11-26
修回日期: 2013-01-28
网络出版日期: 2013-02-21
基金资助
国家"863"计划(2010AA7010213)
Adaptive Square-root Cubature Kalman Filter Algorithm Based on Gaussian Process Regression Models
Received date: 2012-11-26
Revised date: 2013-01-28
Online published: 2013-02-21
Supported by
National High-tech Research and Development Program of China (2010AA7010213)
与传统算法一样,动态系统的参数化模型(含噪声统计特性)未知或不够准确易导致容积卡尔曼滤波(CKF)效果严重下降,甚至滤波结果发散。为此,利用高斯过程回归(GPR)方法对训练数据进行学习,得到动态系统的状态转移GPR模型和量测GPR模型以及噪声统计特性,用以替代或增强原有动态系统模型,并将其融入到平方根容积卡尔曼滤波(SRCKF)中,分别提出了无模型高斯过程SRCKF(MFGP-SRCKF)和模型增强高斯过程SRCKF(MEGP-SRCKF)两种算法。仿真结果表明:这两种新的自适应滤波算法提高了动态系统模型精度,且实时自适应调整噪声的协方差,克服了传统算法滤波性能易受系统模型限制的问题;与MFGP-SRCKF相比,在给定一个不够准确的参数化模型,且有限的训练数据未能遍布估计状态空间的情况下,MEGP-SRCKF具备更高的滤波精度。
何志昆 , 刘光斌 , 赵曦晶 , 刘冬 , 张博 . 基于GPR模型的自适应平方根容积卡尔曼滤波算法[J]. 航空学报, 2013 , 34(9) : 2202 -2211 . DOI: 10.7527/S1000-6893.2013.0118
In many applications, the parametric models of dynamical systems (including the process and measurement of noise statistics) are difficult to obtain or are insufficiently accurate, which results in the serious deterioration or even divergence of the filtering of cubature Kalman filter (CKF). In this paper, the Gaussian process regression (GPR) method is used to learn the training data to obtain the transition and measurement GPR models and their noise statistics of dynamical systems. These GPR models are used to replace or enhance the primary system models and integrate them into the square-root CKF (SRCKF), which yields a model-free Gaussian process SRCKF (MFGP-SRCKF) algorithm and a model-enhanced Gaussian process SRCKF (MEGP-SRCKF) algorithm. Simulation results show that, by improving the accuracy of the models of dynamical systems and adjusting adaptively the noise covariance real-time, the two new adaptive filters alleviate the problem of unknown or insufficiently accurate system models in the classical filters. Meanwhile, in the case that an insufficiently accurate parametric model is given and the limited training data do not fill all over the estimated state space, MEGP-SRCKF can yield higher filtering accuracy than MFGP-SRCKF.
[1] Ho Y C, Lee R C K. A Bayesian approach to problems in stochastic estimation and control. IEEE Transactions on Automatic Control, 1964, 9(4): 333-339.
[2] Stone C J. A course in probability and statistics. Belmont: Duxbury Press, 1996.
[3] Jazwinski A H. Stochastic processes and filtering theory. New York: Academic Press, 1970.
[4] Simon D. Optimal state estimation: Kalman, H∞ and nonlinear approaches. New York: John Wiley & Sons, Inc., 2006.
[5] Bell B M, Cathey F W. The iterated Kalman filter update as a Gauss-Newton method. IEEE Transactions on Automatic Control, 1993, 38(2): 294-297.
[6] Galkowski P J, Islam M A. An alternative derivation of modified gain function of Song and Speyer. IEEE Transactions on Automatic Control, 1991, 36(11): 1323-1326.
[7] Julier S J, Uhlmann J K. Unscented filtering and nonlinear estimation. Proceedings of the IEEE, 2004, 92(3): 401-422.
[8] Ito K, Xiong KQ. Gaussian filters for nonlinear filtering problems. IEEE Transactions on Automatic Control, 2000, 45(5): 910-927.
[9] Arasaratnam I, Haykin S, Elliott R J. Discrete-time nonlinear filtering algorithms using Gauss-Hermite quadrature. Proceedings of the IEEE, 2007, 95(5): 953-977.
[10] Arasaratnam I, Haykin S. Cubature Kalman filters. IEEE Transactions on Automatic Control, 2009, 54(6): 1254- 1269.
[11] Huang J J, Zhong J L, Jiang F. A CKF based spatial alignment of radar and infrared sensors. 2010 IEEE 10th International Conference on Signal Processing, 2010: 2386-2390.
[12] Liu J, Cai B G, Tang T, et al. A CKF based GNSS/INS train integrated positioning method. 2010 International Conference on Mechatronics and Automation, 2010: 1686-1689.
[13] Pesonen H, Piché R. Cubature-based Kalman filters for positioning. 2010 7th Workshop on Positioning Navigation and Communication, 2010: 45-49.
[14] Mu J, Cai Y L, Zhang J M. Square root cubature particle filter. Advanced Materials Research, 2011, 219-220: 727-731.
[15] Rasmussen C E, Williams C K I. Gaussian processes for machine learning. Cambridge: MIT Press, 2006: 7-31.
[16] Ko J, Fox D. GP-BayesFilters: Bayesian filtering using Gaussian process prediction and observation models. Automomous Robots, 2009, 27(1): 75-90.
[17] Gregorcˇicˇ G, Lightbody G. Gaussian processes for modelling of dynamic non-linear systems. Proceedings of the Irish Signals and Systems Conference, 2002: 141-147.
[18] Ni W D, Tan S K, Ng W J, et al. Moving-window GPR for nonlinear dynamic system modeling with dual updating and dual preprocessing. Industrial and Engineering Chemistry Research, 2012, 51(18): 6416-6428.
[19] Ferris B, Haehnel D, Fox D. Gaussian processes for signal strength-based location estimation. Proceedings of the International Conference on Robotics, Science and Systems, 2006.
[20] Bar-Shalom Y, Li X R, Kirubarajan T. Estimation with applications to tracking and navigation. New York: John Wiley & Sons, Inc., 2001.
[21] Middlebrook D L. Bearing-only tracking automation for a single unmanned underwater vehicle. Cambridge: Department of Mechanical Engineering, Massachusetts Institute of Technology, 2009.
/
| 〈 |
|
〉 |
CJA