Evaluation of a Helicopter Model Using Generalized Back Propagation Through Time

  • Krzysztof Fujarewicz
Conference paper
Part of the Advances in Soft Computing book series (AINSC, volume 19)


This paper deals with a problem of evaluation of nonlinear model of the helicopter. The plant is an educational apparatus with two degree of freedom. Whereas the theory of linear identification is well established, there are several possible approaches to identification or evaluation of nonlinear plants. Here so called Generalized Back Propagation Through Time Method (GBPTT) is utilized. This method, derived originally for recurrent neural networks, can be used for any nonlinear dynamical system given as a block diagram. The method is fully structural and, in addition, mnemonic. The results of applying of GBPTT for the helicopter shows that obtained model outperforms the model achieved by separate measurements or estimation of its parameters.


Block Diagram Nonlinear Dynamical System Recurrent Neural Network Sensitivity Model Suboptimal Control 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Ahmed, M. S. (1994) Block partial derivative and its application to neural-net-based direct-model-reference adaptive control, IEE Proc.-Control Theory Appl., 141, 305–314MATHCrossRefGoogle Scholar
  2. 2.
    Baldi, P. (1995) Gradient descent learning algorithm overview: A general dynamical systems perspective, IEEE Trans. Neural Networks, 6, 182–195CrossRefGoogle Scholar
  3. 3.
    Cruz, J. B. (Ed.) (1973) System sensitivity analysis, Benchmark papers in electrical engineering and computer science, Dowden, Hutchinson & Ross, Inc., StroudsburgGoogle Scholar
  4. 4.
    Fujarewicz, K. (1999) Optimal control of nonlinear systems using generalized back propagation through time, Proc. Eighteenth IASTED Conference Modelling, Identification and Control, Innsbruck, Austria, 468–471Google Scholar
  5. 5.
    Fujarewicz, K. (1999) Systems transposed in time and their application to gradient calculation in dynamical systems containing neural nets, Proc. Fourth Conference–Neural Networks and Their Applications, Zakopane, Poland, 395–400Google Scholar
  6. 6.
    Fujarewicz, K. (1999) Control of certain nonlinear systems using neural networks - Ph.D. thesis, Silesian University of TechnologyGoogle Scholar
  7. 7.
    Fujarewicz, K. (2000) Identification of heat exchanger using generalized back propagation through time, Proc. Nineteenth IASTED Conference Modelling, Identification and Control, Innsbruck, Austria, 267–273Google Scholar
  8. 8.
    Fujarewicz, K. (2000) Identification and suboptimal control of heat exchanger using generalized back propagation through time, Archives of Control Sciences, 10(XLVI), No. 3–4, 167–183Google Scholar
  9. 9.
    Kailath, T. (1980) Linear Systems, Prentice-Hall, Inc., Engelwood Cliffs, N.J.Google Scholar
  10. 10.
    Narendra, S. N., Parthasarathy, K. (1991) Gradient methods for the optimization of dynamical systems containing neural networks, IEEE Trans. Neural Networks, 2, 252–262CrossRefGoogle Scholar
  11. 11.
    Pearlmutter, B. A. (1995) Gradient calculations for dynamic recurrent neural networks: A survey, IEEE Trans. Neural Networks, 6, 1212–1228.CrossRefGoogle Scholar
  12. 12.
    Wan, E., Beaufays, F. (1996) Diagrammatic derivation of gradient algorithms for neural networks, Neural Computation, 8, no. 1, 182–201CrossRefGoogle Scholar
  13. 13.
    Werbos, P. J. (1990) Backpropagation through time: what it does and how to do it, Proc. IEEE, 78, 1550–1560CrossRefGoogle Scholar
  14. 14.
    Wierzbicki, A. (1977) Models and sensitivity of control systems, WNT, Warsaw, 1977Google Scholar
  15. 15.
    Williams, R., Zipser, D. (1989) Learning algorithm for continually running fully recurrent neural networks, Neural Computation, 1, 270–280CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Krzysztof Fujarewicz
    • 1
  1. 1.Silesian University of TechnologyGliwicePoland

Personalised recommendations