Training of Neural Network Wiener Models with Recursive Prediction Error Algorithm
Identification of the Wiener system consisting of a linear dynamics in series with a static nonlinearity is considered. A recursive prediction error training algorithm for a recurrent neural network Wiener model is proposed. The gradient of the model output w.r.t. parameters of its linear part is computed with the sensitivity method. The proposed algorithm has superior convergence properties in comparison with gradient methods such as the sensitivity method or the truncated backpropagation through time. Its performance is illustrated with a simulated example of a pneumatic valve.
KeywordsRecurrent Neural Network Gradient Descent Method Nonlinear Element Linear Dynamic Model Pneumatic Valve
Unable to display preview. Download preview PDF.
- 2.Janczak A. (2001) On identification of Wiener systems based on a modified serial-parallel model. Proc. ECC’01, Porto, Portugal. 1852–1857.Google Scholar
- 4.Marciak C., Latawiec K., Rojek R., Oliveira G. H. C. (2001) Adaptive least-squares parameter estimation of OBF-based Wiener models. Proc. 7th IEEE Conf. MMAR’2001, Migdzyzdroje, Poland. 965–969.Google Scholar
- 9.Wigren T. (1994) Convergence analysis of recursive identification algorithms based on the nonlinear Wiener model. IEEE Trans. Automat. Contr. 39, 2191 2206.Google Scholar
- 11.Janczak A. (1995) Identification of a class of nonlinear systems using neural networks. Proc. 2nd Int. Symp. MMAR’95, Migdzyzdroje, Poland. 697–702.Google Scholar
- 12.Janczak A. (1997) Identification of Wiener models using recurrent neural networks. Proc. 4th Int. Symp. MMAR’97, Migdzyzdroje, Poland. 727–732.Google Scholar
- 13.Janczak A. (1998) Recurrent neural network models for identification of Wiener systems. Proc. CESA’98 IMACS Multiconference, Nabeul-Hammamet, Tunisia. 965–970.Google Scholar
- 14.Billings S. A, Jamaluddin H B., Chen S. (1992) Properties of neural networks with application to modelling non-linear dynamical systems. Int. J. Control. 55, 193–224.Google Scholar
- 17.Kaminsky P. G, Bryson A. E., Schmidt S. F. (1971) Discrete square root filtering: A survey of current techniques. IEEE Trans. Autaomat. Control. AC-16, 727–735.Google Scholar