On-line filtering of stochastic variables that are difficult or expensive to directly measure has been widely studied. In this paper a practical algorithm is presented for adaptive state filtering when the underlying nonlinear state equations are partially known. The unknown dynamics are constructively approximated using neural networks. The proposed algorithm is based on the two-step prediction-update approach of the Kalman Filter. The algorithm accounts for the unmodeled nonlinear dynamics and makes no assumptions regarding the system noise statistics. The proposed filter is implemented using static and dynamic feedforward neural networks. Both off-line and on-line learning algorithms are presented for training the filter networks. Two case studies are considered and comparisons with Extended Kalman Filters (EKFs) performed. For one of the case studies, the EKF converges but it results in higher state estimation errors than the equivalent neural filter with on-line learning. For another, more complex case study, the developed EKF does not converge. For both case studies, the off-line trained neural state filters converge quite rapidly and exhibit acceptable performance. On-line training further enhances filter performance, decoupling the eventual filter accuracy from the accuracy of the assumed system model.
Issue Section:
Technical Papers
Keywords:
adaptive filters,
Kalman filters,
filtering theory,
nonlinear equations,
nonlinear dynamical systems,
noise,
feedforward neural nets,
state estimation,
convergence,
Adaptive State Filtering,
Dynamic Neural Networks,
Nonlinear State Filtering,
Multi-Step Prediction,
Extended Kalman Filtering
1.
Kalman
, R. E.
, and Bucy
, R. S.
, 1961
, “New Results in Linear Filtering and Prediction Theory
,” ASME J. Basic Eng.
, 83
, pp. 95
–107
.2.
Gelb, A., 1974, Applied Optimal Estimation, MIT Press, Cambridge, MA.
3.
Jonsson
, G.
, and Palsson
, O. P.
, 1994
, “An Application of Extended Kalman Filtering to Heat Exchanger Models
,” ASME J. Dyn. Syst., Meas., Control
, 116
, pp. 257
–264
.4.
Haykin, S., 1999, Neural Networks: A Comprehensive Foundation, 2nd Edition, Prentice-Hall, Piscataway, NJ.
5.
Lo
, J. T-H.
, 1994
, “Synthetic Approach to Optimal Filtering
,” IEEE Trans. Neural Netw.
, 5
(5
) Sept., pp. 803
–811
.6.
Elanayar
, S.
, and Shin
, Y. C.
, 1994
, “Radial Basis Function Neural Network for Approximation and Estimation of Nonlinear Stochastic Dynamic Systems
,” IEEE Trans. Neural Netw.
, 5
(4
) July, pp. 594
–603
.7.
Annaswamy
, A. M.
, and Yu
, S. H.
, 1996
, “θ-Adaptive Neural Networks: A New Approach to Parameter Estimation
,” IEEE Trans. Neural Netw.
, 7
(4
), pp. 594
–603
.8.
Parisini, T., Alessandri, A., Maggiore, M., and Zoppoli, R., 1997, “On Convergence of Neural Approximate Nonlinear State Estimators,” Proceedings of the 1997 American Control Conference, Vol. 3, June, pp. 1819–1822.
9.
Alessandri
, A.
, Parisini
, T.
, and Zoppoli
, R.
, 1997
, “Neural Approximators for Nonlinear Finite-Memory State Estimation
,” Int. J. Control
, 67
(2
), pp. 275
–301
.10.
Haykin
, S.
, Yee
, P.
, and Derbez
, E.
, 1997
, “Optimum Nonlinear Filtering
,” IEEE Trans. Neural Netw.
, 45
(11
) Nov., pp. 2774
–2786
.11.
Zhu, R., Chai, T., and Shao, C., 1997, “Robust Nonlinear Adaptive Observer Design using Dynamic Recurrent Neural Networks,” Proceedings of the 1997 American Control Conference, Vol. 2, June, pp. 1096–1100.
12.
Habtom, R., and Litz, L., 1997, “Estimation of Unmeasured Inputs using Recurrent Neural Networks and the Extended Kalman Filter,” International Conference on Neural Networks, Vol. 4, pp. 2067–2071.
13.
Dong
, X.
, Qui
, L.
, and Wang
, Z.
, 1997
, “Neural Networks-based Nonlinear Adaptive Filters and On-line Fault Detection
,” Control and Decision
, 12
(1
) Jan., pp. 78
–87
.14.
Lei, J., Guangdong, H., and Jiang, J. P., 1997, “The State Estimation of the CSTR System Based on a Recurrent Neural Network Trained by HGAs,” International Conference on Neural Networks, Vol. 2, pp. 779–782.
15.
Schenker
, B.
, and Agarwal
, M.
, 1998
, “Predictive Control of a Bench-Scale Chemical Reactor Based on Neural-Network Models
,” IEEE Trans. Control Syst. Technol.
, 6
(3
) May, pp. 388
–400
.16.
Stubberud, S. C., and Owen, M., 1998, “Targeted On-line Modeling for an Extended Kalman Filter Using Artificial Neural Networks,” Proceedings of the 1998 American Control Conference, Vol. 3, June, pp. 1852–1856.
17.
Stubberud
, S. C.
, Owen
, M.
, and Lobbia
, R. N.
, 1998
, “Adaptive Extended Kalman Filter Using Artificial Neural Networks
,” International Journal of Smart Engineering System Design
, 1
(3
), pp. 207
–221
.18.
Durovic, Z., and Kovacevic, B., 1998, “Adaptive Filtering using Neural Networks Approach,” Proceedings of the Mediterranean Electrotechnical Conference, Vol. 1, May, pp. 499–503.
19.
Menon, S. K., Parlos, A. G., and Atiya, A. F., 2000, “Nonlinear State Filtering for Fault Diagnosis and Prognosis in Complex Systems Using Recurrent Neural Networks,” 4th Symposium on Fault Detection, Supervision and Safety for Technical Processes, IFAC SAFEPROCESS 2000, June.
20.
Parlos, A. G., Menon, S. K., and Atiya, A. F., 1999, “Adaptive State Estimation Using Dynamic Recurrent Neural Networks,” Proceedings of the International Joint Conference on Neural Networks, June.
21.
Grewal, M. S., and Andrews, A. P., 1993, Kalman Filtering: Theory and Practice, Prentice-Hall, Upper Saddle River, NJ.
22.
Ljung, L., 1999, System Identification: Theory for the User, 2nd Edition, Prentice-Hall, Upper Saddle River, NJ.
23.
Narendra
, K. S.
, and Parthasarathy
, K.
, 1990
, “Identification and Control of Dynamic System Using Neural Networks
,” IEEE Trans. Neural Netw.
, 1
, pp. 4
–27
.24.
Barron
, A. R.
, 1994
, “Approximation and Estimation Bounds for Artificial Neural Networks
,” Journal of Machine Learning
, 14
, pp. 115
–133
.25.
Parlos
, A. G.
, Rais
, O. T.
, and Atiya
, A. F.
, 2000
, “Multi-Step-Ahead Prediction in Complex Systems Using Dynamic Recurrent Neural Networks
,” Neural Networks
, 13
(4–5
), pp. 765
–786
.26.
Parlos
, A. G.
, Chong
, K. T.
, and Atiya
, A.
, 1994
, “Application of the Recurrent Multilayer Perceptron in Modeling Complex Process Dynamics
,” IEEE Trans. Neural Netw.
, 5
, pp. 255
–266
.27.
Williams
, R.
, and Zipser
, D.
, 1989
, “A learning algorithm for continually running fully recurrent neural networks
,” Neural Comput.
, 1
, pp. 270
–280
.28.
Atiya
, A.
, and Parlos
, A.
, 2000
, “New Results on Recurrent Network Training: Unifying the Algorithms and Accelerating Convergence
,” IEEE Trans. Neural Netw.
, 11
, pp. 697
–709
.29.
Geiger, G., 1984, “Fault Identification of a Motor-Pump System using Parameter Estimation and Pattern Classification,” IFAC 9th Triennial World Congress, Dec.
30.
Isermann
, R.
, 1985
, “Process Fault Diagnosis with Parameter Estimation Methods
,” IFAC Digital Computer Applications to Process Control
, Dec., pp.
51
–60
.31.
Franklin, G. F., Powell, J. D., and Workman, M., 1998, Digital Control of Dynamic Systems, 3rd Edition, Addison Wesley Longman, Menlo Park, CA.
32.
Choi, J. I., 1987, Nonlinear Digital Computer Control for the Steam Generator System in a Pressurized Water Reactor Plant, PhD thesis, MIT, Department of Nuclear Engineering.
33.
Menon
, S. K.
, and Parlos
, A. G.
, 1992
, “Gain-Scheduled Nonlinear Control of U-Tube Steam Generator Water Level
,” Nuclear Science and Engineering
, III
(3
), pp. 294
–308
.Copyright © 2002
by ASME
You do not currently have access to this content.
