On-Line Learning of a Time Variant System

  • Fernando Morgado Dias
  • Ana Antunes
  • José Vieira
  • Alexandre Manuel Mota
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4031)


In the present work a sliding window approach for the Levenberg-Marquardt algorithm is used for on-line modelling a time variant system. The system used is a first order cruise control in which a modification is introduced to change the system gain at some point of operation. The initial control of the cruise control is performed by a PI not particularly optimised but enough to keep the system working within the intended range, which is then replaced by an Artificial Neural Network as soon as it is trained, using an Internal Model Controller loop.


Hessian Matrix Trust Region Inverse Model Time Variant System Slide Window Approach 
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.
    Levenberg, K.: A method for the solution of certain problems in least squares. Quart. Appl. Math. 2, 164–168 (1944)MATHMathSciNetGoogle Scholar
  2. 2.
    Marquardt, D.: An algorithm for least -squares estimation of nonlinear parameters. SIAM J. Appl. Math. 11, 431–441 (1963)CrossRefMATHMathSciNetGoogle Scholar
  3. 3.
    Nørgaard, M., Ravn, O., Poulsen, N.K., Hansen, L.K.: Neural Networks for Modelling and Control of Dynamic Systems. Springer, Heidelberg (2000)CrossRefGoogle Scholar
  4. 4.
    Ngia, L.S.H.: System Modeling Using Basis Functions and Application to Echo Cancelation, Ph.D. thesis, Department of Signals and Systems School of Electrical and Computer Engineering, Chalmers University of Technology (2000)Google Scholar
  5. 5.
    Ferreira, P., Faria, E., Ruano, A.: Neural network models in greenhouse air temperature prediction. Neurocomputing 43(1-4), 51–75 (2002)CrossRefMATHGoogle Scholar
  6. 6.
    Dias, F.M., Antunes, A., Vieira, J., Mota, A.M.: Implementing the levenberg-marquardt algorithm on-line: A sliding window approach with early stopping. In: 2nd IFAC Workshop on Advanced Fuzzy/Neural Control (2004)Google Scholar
  7. 7.
    Morgan, N., Bourlard, H.: Generalization and parameter estimation in feedforward nets: Some experiments. In: Touretzsky, D. (ed.) Advances in Neural Information Processing Systems, pp. 630–637. Morgan Kaufmann, San Francisco (1990)Google Scholar
  8. 8.
    Sjöberg, J.: Non-Linear System Identification with Neural Networks, Ph.D. thesis, Dept. of Electrical Engineering, Linköping University, Suécia (1995)Google Scholar
  9. 9.
    Dias, F.M., Antunes, A., Mota, A.: A new hybrid direct/specialized approach for generating inverse neural models. WSEAS Transactions on Systems 3(4), 1521–1529 (2004)Google Scholar
  10. 10.
    Nørgaard, M.: Neural network based system identification toolbox for use with matlab, version 1.1, technical report. Tech. Rep. Technical University of Denmark (1996)Google Scholar
  11. 11.
    Nørgaard, M.: Neural network based control system design toolkit for use with matlab, version 1.1, technical report. Tech. Rep., Technical University of Denmark (1996)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Fernando Morgado Dias
    • 1
  • Ana Antunes
    • 1
  • José Vieira
    • 2
  • Alexandre Manuel Mota
    • 3
  1. 1.Departamento de Engenharia Electrotécnica, Campus do IPS, EstefanilhaEscola Superior de Tecnologia de Setúbal do Instituto Politécnico de SetúbalSetúbalPortugal
  2. 2.Departamento de Engenharia ElectrotécnicaEscola Superior de Tecnologia de Castelo BrancoCastelo BrancoPortugal
  3. 3.Departamento de Electrónica e TelecomunicaçõesUniversidade de AveiroAveiroPortugal

Personalised recommendations