Lamine Thiaw, Kurosh Madani, Rachid Malti, Gustave Sow



Multi-modeling is a recent tool proposed for modeling complex nonlinear systems by the use of a combination of relatively simple set of local models. Due to their simplicity, linear local models are mainly used in such structures. In this work, multi-models having polynomial local models are described and applied in system identification. Estimation of model’s parameters is carried out using least squares algorithms which reduce considerably computation time as compared to iterative algorithms. The proposed methodology is applied to recurrent models implementation. NARMAX and NOE multi-models are implemented and compared to their corresponding neural network implementations. Obtained results show that the proposed recurrent multi-model architectures have many advantages over neural network models.


  1. Bezdec, J. (1973). Fuzzy mathematics in pattern classification. PhD thesis, Applied Math. Center, Cornell University Ithaca.
  2. Bielikova, M. (2005). Recurrent neural network training with the extended kalman filter. IIT. SRC, pages 57- 64.
  3. Boukhris, A., Mourot, G., and Ragot, J. (2000). Nonlinear dynamic system identification: a multiple-model approach. Int. J. of control, 72(7/8):591-604.
  4. Box, G. and Jenkins, G. (1970). Time series analysis, forecasting and control. San Francisco, Holden Day, pages 532-533.
  5. Cheng, Y., Karjala, T., and Himmelblau, D. (1997). Closed loop nonlinear process identification using internal recurrent nets. Neural Networks, 10(3):573-586.
  6. Corradini, A. and Cohen, P. (2002). Multimodal speechgesture interface for hands-free painting on virtual paper using partial recurrent neural networks for gesture recognition. in Proc. of the Int'l Joint Conf. on Neural Networks (IJCNN'02), 3:2293-2298.
  7. Dreyfus, G. (2002). Rseaux de Neurones - Mthodologie et applications. Eyrolles.
  8. Elman, J. (1999). Finding structure in time. Cognitive Science, 14(2):179-211.
  9. Gao, Y. and Foss, A. (2005). Narmax time series model prediction: feed forward and recurrent fuzzy neural network approaches. Fuzzy Sets and Sytems, 150:331- 350.
  10. Gasso, K. (2000). Identification de systmes dynamiques non linaires: approche multi - modle . thse de doctorat de l'INPL.
  11. Huang, B., Rashid, T., and Kechadi, M.-T. (2005). A recurrent neural network recognizer for online recognition of handwritten symbols. ICEIS, 2:27-34.
  12. Huang, B., Rashid, T., and Kechadi, M.-T. (2006). Multicontext recurrent neural network for time series applications. International Journal of Computational Intelligence, 3(1):45-54.
  13. Johansen, T. and Er, M. (1992). Nonlinear local model representation for adaptive systems. In Proc. of the IEEE Conf. on Intelligent Control and Instrumentation, volume 2, pages 677-682, Singapore.
  14. Johansen, T. and Er, M. (1993). Constructing narmax using armax. Int. Journal of Control, 58(5):1125-1153.
  15. Jordan, M. (1986). Attractor dynamics and parallelism in a connectionist sequential machine. In Proceedings of IASTED International Conference of the Cognitive Science Society. (Reprinted in IEEE Tutorials Series, New York: IEEE Publishing Services, 1990), pages 531-546, Englewood Cliffs, NJ: Erlbaum.
  16. Konur, U. and Okatan, A. (2004). Time series prediction using recurrent neural network architectures and time delay neural networks. ENFORMATIKA, pages 1305- 1313.
  17. Li, N., Li, S. Y., and Xi, Y. G. (2004). Multi-model predictive control based on the takagi-sugeno fuzzy models: a case study. Information Sciences, 165:247-263.
  18. Murray-Smith, R. and Johansen, T. (1997). Multiple Model Approaches to Modeling and Control. Taylor and Francis Publishers.
  19. Orjuela, R., Maquin, D., and Ragot, J. (2006). Identification des systmes non linaires par une approche multimodle tats dcoupls. Journes Identification et Modlisation Exprimentale JIME'2006 - 16 et 17 novembre - Poitiers.
  20. Palma, F. D. and Magni, L. (2004). A multimodel structure for model predictive control. Annual Reviews in Control, 28:47-52.
  21. Sjoberg, J., Zhang, Q., Ljung, L., Benveniste, A., Delyon, B., Glorennec, P., Hjalmarsson, H., and Juditsky, A. (1995). Nonlinear black-box modeling in system identification: a unified overview. Automatica 31, 31(12):1691-1724.
  22. Tomasz, J. and Jacek, M. (1997). Neural networks tool for stellar light prediction. In Proc. of the IEEE Aerospace Conference, volume 3, pages 415-422, Snowmass, Colorado, USA.
  23. Vartak, A., Georgiopoulos, M., and Anagnostopoulos, G. (2005). On-line gauss-newton-based learning for fully recurrent neural networks. Nonlinear Analysis, 63:867-876.
  24. Vernieuwe, H., Georgieva, O., Baets, B., Pauwels, V., Verhoest, N., and Troch, F. (2004). Comparison of datadriven takagi-sugeno models of rainfall-discharge dynamics. Journal of Hydrology, XX:1-14.
  25. Werbos, P. (1990). Backpropagation through time: What it does and how to do it. Proceedings of the IEEE, 78(10):1550-1560.
  26. Yang, W. Z., L.H., Y., and L., C. (2005). Narmax model representation and its application to damage detection for multi-layer composites. Composite Structures, 68:109-117.

Paper Citation

in Harvard Style

Thiaw L., Madani K., Malti R. and Sow G. (2007). IMPLEMENTATION OF RECURRENT MULTI-MODELS FOR SYSTEM IDENTIFICATION . In Proceedings of the Fourth International Conference on Informatics in Control, Automation and Robotics - Volume 3: ICINCO, ISBN 978-972-8865-84-9, pages 314-321. DOI: 10.5220/0001615503140321

in Bibtex Style

author={Lamine Thiaw and Kurosh Madani and Rachid Malti and Gustave Sow},
booktitle={Proceedings of the Fourth International Conference on Informatics in Control, Automation and Robotics - Volume 3: ICINCO,},

in EndNote Style

JO - Proceedings of the Fourth International Conference on Informatics in Control, Automation and Robotics - Volume 3: ICINCO,
SN - 978-972-8865-84-9
AU - Thiaw L.
AU - Madani K.
AU - Malti R.
AU - Sow G.
PY - 2007
SP - 314
EP - 321
DO - 10.5220/0001615503140321