Spatial-temporal nonlinear filtering based on hierarchical statistical models. (English) Zbl 1037.62096
Summary: A hierarchical statistical model is made up generically of a data model, a process model, and occasionally a prior model for all the unknown parameters. The process model, known as the state equations in the filtering literature, is where most of the scientist’s physical/chemical/biological knowledge about the problem is used. In the case of a dynamically changing configuration of objects moving through a spatial domain of interest, that knowledge is summarized through equations of motion with random perturbations.
Our interest is in dynamically filtering noisy observations on these objects, where the state equations are nonlinear. Two recent methods of filtering, the Unscented Particle filter (UPF) and the Unscented Kalman filter, are presented and compared to the better known extended Kalman filter. Other sources of nonlinearity arise when we wish to estimate nonlinear functions of the object positions; it is here where the UPF shows its superiority, since optimal estimates and associated variances are straightforward to obtain. The longer computing time needed for the UPF is often not a big issue, with the ever faster processors that are available.
This paper is a review of spatial-temporal nonlinear filtering, and we illustrate it in a command and control setting where the objects are highly mobile weapons, and the nonlinear function of object locations is a two-dimensional surface known as the danger-potential field.
Our interest is in dynamically filtering noisy observations on these objects, where the state equations are nonlinear. Two recent methods of filtering, the Unscented Particle filter (UPF) and the Unscented Kalman filter, are presented and compared to the better known extended Kalman filter. Other sources of nonlinearity arise when we wish to estimate nonlinear functions of the object positions; it is here where the UPF shows its superiority, since optimal estimates and associated variances are straightforward to obtain. The longer computing time needed for the UPF is often not a big issue, with the ever faster processors that are available.
This paper is a review of spatial-temporal nonlinear filtering, and we illustrate it in a command and control setting where the objects are highly mobile weapons, and the nonlinear function of object locations is a two-dimensional surface known as the danger-potential field.
MSC:
| 62M20 | Inference from stochastic processes and prediction |
| 62M30 | Inference from spatial processes |
| 62P99 | Applications of statistics |
Keywords:
battlespace; danger-potential field; Kalman filter; particle filter; resampling; scaled unscented transformation; sequential importance sampler; unscented particle filterReferences:
| [1] | Anderson, B. D. andMoore, J. B. (1979).Optimal Filtering. Prentice-Hall, New Jersey. · Zbl 0688.93058 |
| [2] | Bar-Shalom, Y. andFortmann, T. E. (1988).Tracking and Data Association. Academic Press, Boston. · Zbl 0634.93001 |
| [3] | Bergman, N. (2001). Posterior Cramer-Rao bounds for sequential estimation. In A. Doucet, N. de Freitas, and H. J. Gordon, eds.,Sequential Monte Carlo Methods in Practice, pp. 321–338. Springer Verlag, New York. · Zbl 1056.93569 |
| [4] | Bernardo, J. J. (1979). Expected information as expected utility.Annals of Statistics, 7:686–690. · Zbl 0407.62002 · doi:10.1214/aos/1176344689 |
| [5] | Berzuini, C., Best, N. G., Gilks, W. R. andLarizza, C. (1997). Dynamic conditional independence models and markov chain Monte Carlo methods.Journal of the American Statistical Association, 92:1403–1412. · Zbl 0913.62025 · doi:10.2307/2965410 |
| [6] | Bhattacharya, R. N. andWaymire, E. (1990).Stochastic Processes with Applications. Wiley, New York. |
| [7] | Blake, A., Isard, M., andMacCormick, J. (2001). Statistical models of visual shape and motion. In, A. Doucet, N. de Freitas, and H. J. Gordon, eds.,Sequential Monte Carlo Methods in Practice, pp. 339–357. Springer Verlag, New York. · Zbl 1056.93571 |
| [8] | Breckling, J. (1989).The Analysis of Directional Time Series: Applications to Wind Speed and Direction. Springer-Verlag, Berlin. · Zbl 0698.62093 |
| [9] | Brillinger, D. R., Preisler, H. K., Ager, A. A. andKie, J. G. (2000).The Use of Potential Functions in Modelling Animal Movement. Data Analysis from Statistical Foundations. Nova Science, New York. |
| [10] | Chen, R. andLiu, J. S. (2000). Mixture Kalman filters.Journal of the Royal Statistical Society, Series B, 62:493–508. · Zbl 0953.62100 · doi:10.1111/1467-9868.00246 |
| [11] | Christakos, G. (2000).Modern Spatiotemporal Geostatistics. Oxford University Press, Oxford. |
| [12] | Cohn, S. E. andTodling, R. (1996). Approximate data assimilation schemes for stable and unstable dynamics.Journal of the Meteorological Society of Japan, 74:63–75. |
| [13] | Cressie, N., Wendt, D., Johannesson, G., Mugglin, A., andHrafnkelsson, B. (2002). A spatial-temporal statistical approach to problems in command and control. In B. Bodt, ed.,Proceedings of U.S. Army Conference on Applied Statistics 2000. In press. |
| [14] | Crisan, D. (2001). Particle filters–a theoretical perspective. In A. Doucet, N. de Freitas, and H. J. Gordon, eds.,Sequential Monte Carlo Methods in Practice, pp. 17–41. Springer Verlag, New York. · Zbl 1056.93573 |
| [15] | Doucet, A. (1998). On sequential simulation-based methods for Bayesian filtering. Technical report CUED/F-INFENG/TR 310, Department of Engineering, Cambridge University. |
| [16] | Doucet, A., Gordon, N. J., andKrishnamurthy, V. (2002). Particle filters for state estimation of jump Markov linear systems.IEEE Transactions on Signal Processing, 49:613–624. · doi:10.1109/78.905890 |
| [17] | Downs, T. D. andMardia, K. V. (2002). Circular regression.Biometrika, 89:683–697. · Zbl 1037.62056 · doi:10.1093/biomet/89.3.683 |
| [18] | Evensen, G. (1994). Sequential data assimilation with a nonlinear quasigeostrophic model using Monte Carlo methods to forecast error statistics.Journal of Geophysical Research, 99:10143–10162. · doi:10.1029/94JC00572 |
| [19] | Ferguson, T. S. (1967).Mathematical Statistics: A Decision Theoretic Approach. Academic Press, San Diego. · Zbl 0153.47602 |
| [20] | Gelman, A., Carlin, J. B., Stern, H. S., andRubin, D. (1995).Bayesian Data Analysis. Chapman and Hall, London. · Zbl 1279.62004 |
| [21] | Gordon, N. (1994).Bayesian Methods for Tracking, Ph.D. thesis, Imperial College, University of London. |
| [22] | Gordon, N., Marrs, A., andSalmond, D. (2001). Sequential analysis of nonlinear dynamic systems using particles and mixtures. In W. J. Fitzgerald, R. L. Smith, A. T. Walden, and P. Young, eds.,Nonlinear and Nonstationary Signal Processing. Cambridge University Press, Cambridge. · Zbl 0974.93063 |
| [23] | Grandy, W. T. andSchick, L. H., eds. (1991).Maximum Entropy and Bayesian Methods. Kluwer, Dordrecht. |
| [24] | Gregori, P., van Lieshout, M. N. M., andMateu, J. (2002). Generalized area-interaction point processes. In preparation. |
| [25] | Harvey, A. C. (1989).Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge University Press, Cambridge. |
| [26] | Heemink, A. W. (2000). Modeling and prediction of environmental data in space and time using Kalman filtering. In G. B. M. Heuvelink and M. J. P. M. Lemmens, eds.,Proceedings of the 4th International Symposium on Spatial Accuracy Assessment in Natural Resources and Environmental Sciences, pp. 283–291. Delft University Press, Delft. |
| [27] | Higuchi, T. (1997). Monte Carlo filter using the genetic algorithm operators.Journal of Statistical Computation and Simulation, 59:1–23. · Zbl 0896.62097 · doi:10.1080/00949659708811843 |
| [28] | Irwin, M. E., Cox, N., andKong, A. (1994). Sequential imputation for multilocus linkage analysis.Proceedings of the National Academy of Science, USA, 91:11684–11688. · doi:10.1073/pnas.91.24.11684 |
| [29] | Jammalamadaka, S. R. andSenGupta, A. (2001).Topics in Circular Statistics, World Scientific, Singapore. |
| [30] | Jaynes, E. T. (1957). Information theory and statistical mechanics, II.Physics Review, 108:171–190. · Zbl 0084.43701 · doi:10.1103/PhysRev.108.171 |
| [31] | Julier, S. J. (1999). The scaled unscented transformation. Manuscript-http://citeseer.nj.nec.com/julier99scaled.html. |
| [32] | Julier, S. J. andUhlmann, J. K. (1997). A new extension of the Kalman filter to nonlinear systems. In I. Kadar, ed.,Signal Processing, Sensor Fusion, and Target Recognition VI, SPIE Proceedings, vol. 3068, pp. 182–193. SPIE, Bellingham, WA. |
| [33] | Kent, J. T. (1978). Time reversible diffusions.Advances in Applied Probability, 10:819–835. · Zbl 0387.60089 · doi:10.2307/1426661 |
| [34] | Kitagawa, G. (1996). Monte Carlo filter and smoother for non-Gaussian nonlinear state space models.Journal of Computational and Graphical Statistics, 5:1–25. · doi:10.2307/1390750 |
| [35] | Kong, A., Liu, J. S., andWong, W. H. (1994). Sequential imputations and Bayesian missing data problems.Journal of the American Statistical Association, 89:278–288. · Zbl 0800.62166 · doi:10.2307/2291224 |
| [36] | Liu, J. S. (2001).Monte Carlo Strategies in Scientific Computing. Springer Verlag, New York. · Zbl 0991.65001 |
| [37] | Liu, J. S. andChen, R. (1998). Sequential Monte Carlo methods for dynamic systems.Journal of the American Statistical Association, 93:1032–1044. · Zbl 1064.65500 · doi:10.2307/2669847 |
| [38] | Liu, J. S., Chen, R., andLogvinenko, T. (2001). A theoretical framework for sequential importance sampling with resampling. In A. Doucet, N. de Freitas, and H. J. Gordon, eds.,Sequential Monte Carlo Methods in Practice, pp. 17–41, Springer Verlag, New York. · Zbl 1056.93584 |
| [39] | Mardia, K. V. andJupp, P. E. (2000).Directional Statistics, Wiley, New York. |
| [40] | Nelson, E. (1967).Dynamical Theories of brownian Motion. Princeton University Press, Princeton. · Zbl 0165.58502 |
| [41] | Openshaw, S. (1994).Two Exploratory Space-Time-Attribute Pattern Analysers Relevant to GIS. Ed. Stewart Fotheringham and Peter Rogerson. Bristol: Taylor & Francis Inc., pp. 83–104. |
| [42] | Pitt, M. K. andShephard, N. (1999). Filtering via simulation: Auxiliary particle filtersJournal of the American Statistical Association, 94:590–599. · Zbl 1072.62639 · doi:10.2307/2670179 |
| [43] | Ramachandra, K. V. (2000).Kalman Filtering Techniques for Radar Tracking. Marcel Dekker, New York. |
| [44] | Singh, H., Hnizdo, V., andDemchuk, E. (2002). Probabilistic model for two dependent circular variables.Biometrika, 89:719–723. · Zbl 1037.62003 · doi:10.1093/biomet/89.3.719 |
| [45] | Smith, A. F. M. andRoberts, G. (1993). Bayesian computation via the Gibbs sampler and related Markov chain Monte Carlo methods (with discussion).Journal of the Royal Statistical Society, Series B, 55:3–102. |
| [46] | van der Merwe, R., Doucet, A., de Freitas, N., andWan, E. (2000). The unscented particle filter. Technical report CUED/F-INFENG/TR 380, Department of Engineering, Cambridge University. |
| [47] | van der Merwe, R., Doucet, A., de Freitas, N., andWan, E. (2001). The unscented particle filter. In T. K. Leen, T. G. Dietterich, and V. Tresp, eds.,Advances in Neural Information Processing Systems (NIPS13). MIT Press, Cambridge. |
| [48] | Wendt, D., Cressie, N., andIrwin, M. E. (2002). Waypoint analysis for command and control. Technical Report 691, Department of Statistics, The Ohio State University. · Zbl 1075.62104 |
| [49] | West, M. andHarrison, J. (1997).Bayesian Forecasting and Dynamics Models. Springer-Verlag, New York, 2nd ed. |
| [50] | West, M., Harrison, P. J., andMigon, H. S. (1985). Dynamic generalized linear models and Bayesian forecasting (with discussion).Journal of the American Statistical Association, 80:73–96. · Zbl 0568.62032 · doi:10.2307/2288042 |
| [51] | Zhang, J. L. andLiu, J. S. (2002). A new sequential importance sampling method and its application to the 2D HP model.Journal of Chemical Physics, 93:1032–1043. |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
