Kalman Filter and Its Modern Extensions for the Continuous-Time Nonlinear Filtering Problem

[+] Author and Article Information
Amirhossein Taghvaei

Department of Mechanical Science
and Engineering,
University of Illinois at Urbana-Champaign,
Urbana, IL 61801
e-mail: taghvae2@illinois.edu

Jana de Wiljes

Institut für Mathematik,
Universität Potsdam,
Potsdam D-14476, Germany
e-mail: wiljes@uni-potsdam.de

Prashant G. Mehta

Department of Mechanical Science
and Engineering,
University of Illinois at Urbana-Champaign,
Urbana, IL 61801
e-mail: mehtapg@illinois.edu

Sebastian Reich

Institut für Mathematik,
Universität Potsdam,
Potsdam D-14476, Germany;
Department of Mathematics and Statistics,
University of Reading,
Reading RG6 6AX, UK
e-mail: sereich@uni-potsdam.de

Contributed by the Dynamic Systems Division of ASME for publication in the JOURNAL OF DYNAMIC SYSTEMS, MEASUREMENT, AND CONTROL. Manuscript received February 14, 2017; final manuscript received July 31, 2017; published online November 8, 2017. Assoc. Editor: Puneet Singla.

J. Dyn. Sys., Meas., Control 140(3), 030904 (Nov 08, 2017) (11 pages) Paper No: DS-17-1087; doi: 10.1115/1.4037780 History: Received February 14, 2017; Revised July 31, 2017

This paper is concerned with the filtering problem in continuous time. Three algorithmic solution approaches for this problem are reviewed: (i) the classical Kalman–Bucy filter, which provides an exact solution for the linear Gaussian problem; (ii) the ensemble Kalman–Bucy filter (EnKBF), which is an approximate filter and represents an extension of the Kalman–Bucy filter to nonlinear problems; and (iii) the feedback particle filter (FPF), which represents an extension of the EnKBF and furthermore provides for a consistent solution in the general nonlinear, non-Gaussian case. The common feature of the three algorithms is the gain times error formula to implement the update step (to account for conditioning due to the observations) in the filter. In contrast to the commonly used sequential Monte Carlo methods, the EnKBF and FPF avoid the resampling of the particles in the importance sampling update step. Moreover, the feedback control structure provides for error correction potentially leading to smaller simulation variance and improved stability properties. The paper also discusses the issue of nonuniqueness of the filter update formula and formulates a novel approximation algorithm based on ideas from optimal transport and coupling of measures. Performance of this and other algorithms is illustrated for a numerical example.

Copyright © 2018 by ASME
Your Session has timed out. Please sign back in to continue.


Kalnay, E. , 2002, Atmospheric Modeling, Data Assimilation and Predictability, Cambridge University Press, Cambridge, UK. [CrossRef]
Oliver, D. , Reynolds, A. , and Liu, N. , 2008, Inverse Theory for Petroleum Reservoir Characterization and History Matching, Cambridge University Press, Cambridge, UK. [CrossRef]
Kitanidis, P. , 1995, “ Quasi-Linear Geostatistical Theory for Inversion,” Water Resour. Res., 31(10), pp. 2411–2419. [CrossRef]
Burgers, G. , van Leeuwen, P. , and Evensen, G. , 1998, “ On the Analysis Scheme in the Ensemble Kalman Filter,” Mon. Weather Rev., 126(6), pp. 1719–1724. [CrossRef]
Houtekamer, P. , and Mitchell, H. , 2001, “ A Sequential Ensemble Kalman Filter for Atmospheric Data Assimilation,” Mon. Weather Rev., 129(1), pp. 123–136. [CrossRef]
Julier, S. J. , and Uhlmann, J. K. , 1997, “ A New Extension of the Kalman Filter to Nonlinear Systems,” Proc. SPIE, 3068, pp. 182–193.
Reich, S. , and Cotter, C. , 2015, Probabilistic Forecasting and Bayesian Data Assimilation, Cambridge University Press, Cambridge, UK. [CrossRef]
Daum, F. , Huang, J. , and Noushin, A. , 2010, “ Exact Particle Flow for Nonlinear Filters,” Proc. SPIE, 7697, p. 769704.
Daum, F. , Huang, J. , and Noushin, A. , 2017, “ Generalized Gromov Method for Stochastic Particle Flow Filters,” Proc. SPIE, 10200, p. 102000I.
Ma, R. , and Coleman, T. P. , 2011, “ Generalizing the Posterior Matching Scheme to Higher Dimensions Via Optimal Transportation,” IEEE 49th Annual Allerton Conference on Communication, Control, and Computing, Monticello, IL, Sept. 28–30, pp. 96–102.
El Moselhy, T. A. , and Marzouk, Y. M. , 2012, “ Bayesian Inference With Optimal Maps,” J. Comput. Phys., 231(23), pp. 7815–7850. [CrossRef]
Heng, J. , Doucet, A. , and Pokern, Y. , 2015, “ Gibbs Flow for Approximate Transport With Applications to Bayesian Computation,” e-print arXiv:1509.08787. https://arxiv.org/abs/1509.08787
Yang, T. , Blom, H. , and Mehta, P. G. , 2014, “ The Continuous-Discrete Time Feedback Particle Filter,” American Control Conference (ACC), Portland, OR, June 4–6, pp. 648–653.
Crisan, D. , and Xiong, J. , 2007, “ Numerical Solutions for a Class of SPDEs Over Bounded Domains,” ESAIM Proc., 19, pp. 121–125. [CrossRef]
Crisan, D. , and Xiong, J. , 2010, “ Approximate McKean–Vlasov Representations for a Class of SPDEs,” Stochastics, 82(1), pp. 53–68.
Yang, T. , Mehta, P. G. , and Meyn, S. P. , 2011, “ Feedback Particle Filter With Mean-Field Coupling,” 50th IEEE Conference on Decision and Control and European Control Conference (CDC-ECC), Orlando, FL, Dec. 12–15, pp. 7909–7916.
Yang, T. , Mehta, P. G. , and Meyn, S. P. , 2013, “ Feedback Particle Filter,” IEEE Trans. Autom. Control, 58(10), pp. 2465–2480. [CrossRef]
Blom, H. A. P. , 2012, “ The Continuous Time Roots of the Interacting Multiple Model Filter,” 51st IEEE Conference on Decision and Control (CDC), Maui, HI, Dec. 10–13, pp. 6015–6021.
Bar-Shalom, Y. , Daum, F. , and Huang, J. , 2009, “ The Probabilistic Data Association Filter,” IEEE Control Syst. Mag., 29(6), pp. 82–100. [CrossRef]
Kalman, R. E. , 1960, “ A New Approach to Linear Filtering and Prediction Problems,” ASME J. Basic Eng., 82(1), pp. 35–45. [CrossRef]
Law, K. , Stuart, A. , and Zygalakis, K. , 2015, Data Assimilation: A Mathematical Introduction, Springer-Verlag, New York. [CrossRef]
Reich, S. , 2011, “ A Dynamical Systems Framework for Intermittent Data Assimilation,” BIT Numer. Anal., 51(1), pp. 235–249. [CrossRef]
Bergemann, K. , and Reich, S. , 2012, “ An Ensemble Kalman-Bucy Filter for Continuous Data Assimilation,” Meteorol. Z., 21(3), pp. 213–219. [CrossRef]
de Wiljes, J. , Reich, S. , and Stannat, W. , 2016, “ Long-Time Stability and Accuracy of the Ensemble Kalman-Bucy Filter for Fully Observed Processes and Small Measurement Noise,” University Potsdam, Potsdam, Germany, Technical Report. https://www.uni-potsdam.de/de/sfb1294/thema2/research-area-a/a02-long-time-stability-and-accuracy-of-ensemble-transform-filter-algorithms.html
Del Moral, P. , Kurtzmann, A. , and Tugaut, J. , 2016, “ On the Stability and the Uniform Propagation of Chaos of Extended Ensemble Kalman-Bucy Filters,” INRIA Bordeaux Research Center, Talence, France, Technical Report. https://hal.archives-ouvertes.fr/hal-01337716
Kelly, D. , Law, K. , and Stuart, A. , 2014, “ Well-Posedness and Accuracy of the Ensemble Kalman Filter in Discrete and Continuous Time,” Nonlinearity, 27(10), p. 2579. [CrossRef]
Kelly, D. , and Stuart, A. M. , 2016, “ Ergodicity and Accuracy of Optimal Particle Filters for Bayesian Data Assimilation,” e-print arXiv:1611.08761. https://arxiv.org/abs/1611.08761
Kelly, D. , Majda, A. J. , and Tong, X. T. , 2015, “ Concrete Ensemble Kalman Filters With Rigorous Catastrophic Filter Divergence,” Proc. Natl. Acad. Sci. U.S.A., 112(34), pp. 10589–10594. [CrossRef] [PubMed]
Oksendal, B. , 2013, Stochastic Differential Equations: An Introduction With Applications, Springer Science & Business Media, Berlin. [PubMed] [PubMed]
Zhang, C. , Taghvaei, A. , and Mehta, P. G. , 2016, “ Feedback Particle Filter on Matrix Lie Groups,” American Control Conference (ACC), Boston, MA, July 6–8, pp. 2723–2728.
Yang, T. , Laugesen, R. S. , Mehta, P. G. , and Meyn, S. P. , 2016, “ Multivariable Feedback Particle Filter,” Automatica, 71, pp. 10–23. [CrossRef]
Villani, C. , 2003, Topics in Optimal Transportation, Vol. 58, American Mathematical Society, Providence, RI. [CrossRef] [PubMed] [PubMed]
Evans, L. C. , 1997, “ Partial Differential Equations and Monge-Kantorovich Mass Transfer,” Current Developments in Mathematics, Boston, MA, pp. 65–126. [CrossRef]
Laugesen, R. S. , Mehta, P. G. , Meyn, S. P. , and Raginsky, M. , 2015, “ Poisson’s Equation in Nonlinear Filtering,” SIAM J. Control Optim., 53(1), pp. 501–525. [CrossRef]
Stano, P. M. , Tilton, A. K. , and Babuška, R. , 2014, “ Estimation of the Soil-Dependent Time-Varying Parameters of the Hopper Sedimentation Model: The FPF Versus the BPF,” Control Eng. Pract., 24, pp. 67–78. [CrossRef]
Tilton, A. K. , Ghiotto, S. , and Mehta, P. G. , 2013, “ A Comparative Study of Nonlinear Filtering Techniques,” 16th International Conference on Information Fusion (FUSION), Istanbul, Turkey, July 9–12, pp. 1827–1834. http://ieeexplore.ieee.org/document/6641226/
Berntorp, K. , 2015, “ Feedback Particle Filter: Application and Evaluation,” 18th International Conference on Information Fusion (FUSION), Washington, DC, July 6–9, pp. 1633–1640. http://ieeexplore.ieee.org/document/7266752/
Berntorp, K. , and Grover, P. , 2016, “ Data-Driven Gain Computation in the Feedback Particle Filter,” American Control Conference (ACC), Boston, MA, July 6–8, pp. 2711–2716.
Coifman, R. R. , and Lafon, S. , 2006, “ Diffusion Maps,” Appl. Comput. Harmonic Anal., 21(1), pp. 5–30. [CrossRef]
Hein, M. , Audibert, J. , and Luxburg, U. , 2007, “ Graph Laplacians and Their Convergence on Random Neighborhood Graphs,” J. Mach. Learn. Res., 8, pp. 1325–1368. http://www.jmlr.org/papers/volume8/hein07a/hein07a.pdf
Taghvaei, A. , Mehta, P. G. , and Meyn, S. P. , 2017, “ Error Estimates for the Kernel Gain Function Approximation in the Feedback Particle Filter,” American Control Conference (ACC), Seattle, WA, May 24–26, pp. 4576–4582.
Xiong, J. , 2008, An Introduction to Stochastic Filtering Theory (Oxford Graduate Texts in Mathematics), Vol. 18, Oxford University Press, Oxford, UK.


Grahic Jump Location
Fig. 1

Innovation error-based feedback structure for the (a) Kalman filter and (b) nonlinear FPF

Grahic Jump Location
Fig. 2

The exact solution to the Poisson equation using the formula (13). The density ρ is the sum of two Gaussians N(−1,σ2) and N(+1,σ2), and h(x) = x. The density is depicted as the shaded curve in the background.

Grahic Jump Location
Fig. 3

Constant gain approximation in the FPF

Grahic Jump Location
Fig. 4

Comparison of the gain function approximations obtained using Galerkin (a), kernel (c), and the optimal coupling (e) algorithms. The exact gain function is depicted as a solid line and the density ρ is depicted as a shaded region in the background. (b), (d), and (e) Depict the Monte Carlo estimate of the empirical error (26) as a function of the number of particles N. The Monte Carlo estimate is obtained by averaging the empirical error over 100 simulations. (a) Galerkin gain approximation for N = 100, (b) Galerkin gain approx. error as a function of N, (c) Kernel-based gain approximation for N = 100, (d) Kernel-based gain approx. error as a function of N, (e) optimal coupling gain approximation for N = 100, and (f) optimal coupling gain approx. error as a function of N.

Grahic Jump Location
Fig. 5

Comparison of the Monte Carlo estimate of the empirical error (26) as a function of the parameter ε. The number of particles N = 200.



Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In