Application of an Extended Kalman Filter for On-line Identification with Recurrent Neural Networks (original) (raw)

Abstract

Given their good approximation capabilities for reasonable non-linear systems artificial neural networks have attracted an increasing interest in a number of fields such as system identification and filtering. The main goal of this work is to emphasise the potential benefits of non-linear state-space neural networks for realtime identification with an extended Kalman filter. Experimental results from a laboratory heating system confirm the feasibility of this methodology.

Loading...

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

References (13)

  1. K. Narendra and K. Parthasarathy, Identification and Control of Dynamical Systems Using Neural Networks, IEEE Trans. On Neural Networks, 1, 1, pp. 4- 27, 1990.
  2. K. Hunt, D. Sbarbaro, R. Zbikowski and P. Gawthrop, Neural Networks for Control Systems -A Survey, Automatica, 28, 6, pp. 1083-1112, 1992.
  3. K. Hornik, M. Stinchombe and H. White, Multilayer feedforward networks are universal approximators, Neural Networks, 2, pp 359-366, 1989.
  4. K. Funashi, On the approximate realization of continuous mapping by neural networks, Neural Networks, 2, pp. 183-192, 1989.
  5. L. Jin, P. Nikiforuk and M. Gupta, Approximation of Discrete-Time State-Space Trajectories Using Dynamic Recurrent Neural Networks, IEEE Trans. Automatic Control, 40, 7, pp. 1266-1270, 1995.
  6. I. Rivals and L. Personnaz, Black Box Modelling with State-space Neural Networks, Neural Adaptive Control Technology, R. Zbikowski and K. Hunt eds., World Scientific, pp. 237-264, 1996.
  7. R. Babuska, J. Sousa and H. Verbruggen, Predictive Control of Nonlinear Systems based on Fuzzy and Neural Models, Proc. ECC'99, pp. CA-12-4, 1999.
  8. J. Hopfield, Neurons with graded response have collective computational properties like those of two-state neurons, Proc. National Academic of Sciences USA, 81, pp. 3088-3092, 1984.
  9. J. Elman, Finding Structures in Time, Cognitive Science, 14, pp. 179-211, 1990.
  10. A. Delgado, C. Kambhampi and K. Warwick, Dynamic recurrent neural network for system identification and control, IEE Proc.-Control Theory Appl., 142, 4, pp. 307-314, 1995.
  11. S. Singhal and L. Wu, Training multilayer perceptrons with the extended Kalman filter, In advances in Neural Information Processing Systems 1, Morgan Kauffman, pp. 133-140, 1989.
  12. E. Plumer, Training Neural Networks Using Sequential-Update Forms of The Extended Kalman Filter, LA-UR-95-422 C.2, Los Alamos National Laboratory, 1995.
  13. E. Wan and A. Nelson, Dual EKF Methods, In Kalman Filtering and Neural Networks, ed. Simon Haykin, Wiley Publishing, 2000.