基于最大熵方法的鲁棒自适应滤波及其应用

doi:10.3969/j.issn.1001-506X.2020.03.022

摘要/Abstract

摘要：

针对过程噪声和量测噪声受到脉冲噪声影响而呈现非高斯分布,且噪声统计特性不精确从而导致估计精度下降的问题,提出一种基于最大熵方法的变分贝叶斯自适应卡尔曼滤波(maximum correntropy variational Bayes adaptive Kalman filter, MCVBAKF)算法,并将其应用于捷联惯导系统(strapdown inertial navigation system, SINS)/全球卫星定位系统(global positioning system, GPS)组合导航系统。首先,使用最大熵鲁棒滤波方法对由脉冲噪声产生的野值问题进行处理;然后,通过改进的变分贝叶斯自适应方法进行后验更新,估计噪声,收敛所估参数的估计值;最后,进行了仿真对比。结果表明, MCVBAKF在复杂环境下可以有效提升滤波精度和稳定性。

关键词: 最大熵方法, 鲁棒滤波, 自适应滤波, 组合导航

Abstract:

Aiming at the problem that the estimation accuracy is degraded as the process noise and measurement noise are affected by impulse noise and becoming non-Gaussian distribution and the noise statistical characteristics is inaccurate, a maximum correntropy variational Bayes adaptive Kalman filter (MCVBAKF) is proposed and applied to the strapdown inertial navigation system./global positioning system (SINS/GPS) integrated navigation system. Firstly, the maximum entropy robust filtering method is used to deal with the outlier problem caused by the impulse noise, and is effectively dealt with. Then, a posterior update is made with the improved variational Bayesian adaptive method, that estimate the noise and converge the estimated value. Finally, the end of this paper, the simulation results show that the MCVBAKF can effectively improve the filtering accuracy and stability in complex environment.

Key words: maximum correntropy method, robust filtering, adaptive filtering, integrated navigation

中图分类号:

V249.31

罗凯鑫, 吴美平, 范颖. 基于最大熵方法的鲁棒自适应滤波及其应用[J]. 系统工程与电子技术, 2020, 42(3): 667-673.

Kaixin LUO, Meiping WU, Ying FAN. Robust adaptive filtering based on maximum entropy method and its application[J]. Systems Engineering and Electronics, 2020, 42(3): 667-673.

图/表 12

图1

表1

图2

表2

图3

表3

图4

表4

图5

表5

图6

表6

参考文献 31

1	秦永元, 张洪钺, 王叔华. 卡尔曼滤波与组合导航原理[M]. 西安: 西北工业大学出版社, 2012.
	QIN Y Y , ZHANG H Y , WANG S H . Kalman filter and integrated navigation principle[M]. Xi'an: Northwest University of Technology Press, 2012.
2	KALMAN R E . A new approach to linear filtering and prediction problems[J]. Journal of Basic Engineering Transactions, 1960, 82 (1): 35- 45.
3	HUANG Y L , ZHANG Y G . A new process uncertainty robust student's t based Kalman filter for SINS_GPS integration[J]. IEEE Access, 2017, 5, 14391- 14404.
4	CHEN B D , LIU X , ZHAO H . Maximum correntropy Kalman filter[J]. Automatica, 2017, 76, 70- 77.
5	MOHAMED A H , SCHWARZ K P . Adaptive Kalman filtering for INS/GPS[J]. Journal of Geodesy, 1999, 73 (4): 193- 203.
6	HIDE C, MOORE T, SMITH M.Adaptive Kalman filtering algorithms for integrating GPS and low cost INS[C]//Proc.of the Position Location and Navigation Symposium, 2004: 227-233.
7	YU M J . INS/GPS integration system using adaptive filter for estimating measurement noise variance[J]. IEEE Trans.on Aerospace and Electronic Systems, 2012, 48 (2): 1786- 1792.
8	CHAMBERS J . A robust Gaussian approximate fixed-interval smoother for nonlinear systems with heavy-tailed process and measurement noises[J]. IEEE Signal Processing Letters, 2016, 23 (4): 468- 472.
9	LIU X , CHEN B D , ZHAO H Q . Maximum correntropy Kalman filter with state constraints[J]. IEEE Access, 2017, 5, 25846- 25853.
10	LIU W , POKHAREL P , PRINCIPE J . Correntropy: properties and applications in non-Gaussian signal processing[J]. IEEE Trans.on Signal Processing, 2007, 55 (11): 5286- 5298.
11	WANG L S, GAO X W, YIN L J.Maximum correntropy criterion based robust Kalman filter[C]//Proc.of the China Satellite Navigation Conference, 2018: 491-500.
12	LIU X , QU H , ZHAO J H . Maximum correntropy unscented Kalman filter for spacecraft relative state estimation[J]. Sensors, 2016, 16 (9): 1530.
13	WANG G Q, LI N, ZHANG Y G.Maximum correntropy unscented Kalman and information filters for non-Gaussian measurement noise[J]. Journal of the Franklin Institute, 2017. https://doi.org/10.1016/j.jfranklin.2017.10.023.
14	LI T C, PRIETO J, CORCHADO J M.A short revisit of nonlinear Gaussian filters: state-of-the-art and some concerns[C]//Proc.of the IEEE International Conference on Ubiquitous Wireless Broadband, 2016: 1-4.
15	LI T C , SU J , LIU W . Approximate Gaussian conjugacy: parametric recursive filtering under nonlinearity, multimodality, uncertainty, and constraint, and beyond[J]. Frontiers of Information Technology & Electronic Engineering, 2017, 18 (12): 1913- 1939.
16	FAN Y , ZHANG Y G , WANG G Q , et al. Maximum correntropy based unscented particle filter for cooperative navigation with heavy-tailed measurement noises[J]. Sensors, 2018, 18 (10): 3183.
17	HUANG Y L , ZHANG Y G , WU Z M , et al. A novel adaptive Kalman filter with inaccurate process and measurement noise covariance matrices[J]. IEEE Trans.on Automatic Control, 2018, 63 (2): 594- 601.
18	LI W , SUN S , JIA Y , et al. Robust unscented Kalman filter with adaptation of process and measurement noise covariances[J]. Digital Signal Processing, 2016, 48, 93- 103.
19	JINDICH D , MIROSLAV I , STRAKA O . Methods for estimating state and measurement noise covariance matrices:aspects and vomparison[J]. IFAC Proceedings Volumes, 2009, 42 (10): 372- 377.
20	WANG Y D , ZHENG W , SUN S M . Robust information filter based on maximum correntropy criterion[J]. Journal of Guidance Control & Dynamics, 2016, 39 (5): 1- 6.
21	ZHAO J , NETTO M , MILI L . A robust iterated extended Kalman filter for power system dynamic state estimation[J]. IEEE Trans.on Power Systems, 2017, 32 (4): 3205- 3216.
22	MATISKO P , HAYLENA V . Noise covariance estimation for Kalman filter tuning using Bayesian approach and Monte Carlo[J]. International Journal of Adaptive Control and Signal Processing, 2013, 27 (11): 957- 973.
23	TZIKAS D , LIKAS A , GALATSANOS N . The variational approximation for Bayesian inference[J]. IEEE Signal Process, 2008, 25 (6): 131- 146.
24	邓正隆. 惯性导航原理[M]. 哈尔滨: 哈尔滨工业大学出版社, 1994.
	DENG Z L . Principle of inertial navigation[M]. Harbin: Harbin University of Technology Press, 1994.
25	秦永元. 惯性导航[M]. 北京: 科学出版社, 2006.
	QIN Y Y . Inertial navigation[M]. Beijing: Science Press, 2006.
26	沈忱, 徐定杰, 沈锋. GPS/INS组合导航的变分贝叶斯自适应卡尔曼滤波[J]. 哈尔滨:哈尔滨工业大学学报, 2014, 46 (5): 59- 65.
	SHEN S , XU D J , SHEN F . Variational bayesian adaptive Kalman filtering for GPS/INS integrated navigation[J]. Harbin: Journal of Harbin Institute of Technology, 2014, 46 (5): 59- 65.
27	WANG G Q , ZHANG Y G , WANG X D . Iterated maximum correntropy unscented Kalman filters for non-Gaussian systems[J]. Signal Process, 2019, 163, 87- 94.
28	刘军, 刘克诚, 田甜, 等. 基于自适应卡尔曼滤波算法的紧组合导航系统的研究[J]. 电子测量技术, 2019, 42 (5): 52- 55.
	LIU J , LIU K C , TIAN T , et al. Research on tight integrated navigation system based on adaptive Kalman filter algorithm[J]. Electronic Measurement Technology, 2019, 42 (5): 52- 55.
29	ZHANG Y , JIA G , LI L , et al. A novel adaptive Kalman filter with colored measurement noise[J]. IEEE Access, 2018, 6, 74569- 74578.
30	LIU X , CHEN B D , XU B , et al. Maximum correntropy unscented filter[J]. International Journal of Systems Science, 2017, 48 (8): 1607- 1615.
31	TSENG C , LIN S , JWO D . Robust huber-based cubature kalman filter for GPS navigation processing[J]. Journal of Navigation, 2016, 70 (3): 1- 20.

算法	姿态误差方差/(′)
算法	纵摇角	横摇角	航向角
KF	2.354	4.861	16.13
MCKF	0.736	3.806	6.474
MCVBAKF	0.430	3.415	5.076

算法	速度误差方差/(m/s)
算法	东向	北向
KF	2.051	4.655
MCKF	2.000	1.578
MCVBAKF	1.337	1.554

算法	位置误差方差/(′)
算法	经度	纬度
KF	0.673	1.062
MCKF	0.697	0.404
MCVBAKF	0.233	0.177

算法	姿态误差方差/(′)
算法	纵摇角	横摇角	航向角
KF	1.651	3.934	7.712
MCKF	0.573	3.652	6.515
MCVBAKF	0.545	3.257	4.512

算法	速度误差方差/(m/s)
算法	东向	北向
KF	2.234	2.509
MCKF	1.780	1.841
MCVBAKF	0.991	1.735