Abstract
This article presents a new method for system identification based on dynamic neural networks using prior knowledge. A discrete chart is derived from a given signal flow chart. This discrete chart is implemented in a dynamic neural network model. The weights of the model correspond to physical parameters of the real system. Nonlinear parts of the signal flow chart are represented by nonlinear subparts of the neural network. An optimization algorithm trains the weights of the dynamic neural network model. The proposed identification approach is tested with a nonlinear two mass system.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
M.J. Brache, Identification of dynamic systems using previous knowledge, Diploma Theses, Lehrstuhl für Elektrische Antriebssysteme, Technische Universität München, 2008
C. Endisch, C. Hackl, D. Schröder, Optimal brain surgeon for general dynamic neural networks, in Lecture Notes in Artificial Intelligence (LNAI) 4874 (Springer, Berlin, 2007), pp. 15–28
C. Endisch, C. Hackl, D. Schröder, System identification with general dynamic neural networks and network pruning. Int. J. Computat. Intel. 4(3), 187–195 (2008)
C. Endisch, P. Stolze, C. Hackl, D. Schröder, Comments on backpropagation algorithms for a broad class of dynamic networks. IEEE Trans. Neural Network 20(3), 540–541 (2009)
M. Hagan, B.M. Mohammed, Training feedforward networks with the marquardt algorithm. IEEE Trans. Neural Networks 5(6), 989–993 (1994)
C. Hintz, B. Angerer, D. Schrder, Online identification of mechatronic system with structured recurrent neural networks. Proceedings of the IEEE-ISIE 2002, L’Aquila, Italy, pp. 288–293
C. Hintz, Identifikation nichtlinearer mechatronischer Systeme mit strukturierten rekurrenten Netzen. Dissertation, Lehrstuhl fr Elektrische Antriebssysteme, Technische Universität München, 2003
O. De Jesus, M. Hagan. Backpropagation Algorithms Through time for a general class of recurrent network. IEEE International Joint Conference Neural Network, Washington, 2001, pp. 2638–2643
O. De Jesus, M. Hagan, Forward perturbation algorithm for a general class of recurrent network. IEEE International Joint Conference Neural Network, Washington, 2001, pp. 2626–2631
O. De Jesus, Training general dynamic neural networks. Ph.D. dissertation, Oklahoma State University, Stillwater, OK, 2002
O. De Jesus, M. Hagan backpropagation algorithms for a broad class of dynamic networks. IEEE Trans Neural Networks 18(1), 14–27 (2007)
O. Nelles, in Nonlinear System Identification (Springer, Berlin, 2001).
D. Schröder, in Elektrische Antriebe – Regelung von Antriebssystemen, 2nd edn. (Springer, Berlin, 2001)
P.J. Werbos, Backpropagation through time: what it is and how to do it. Proc. IEEE 78(10), 1550–1560 (1990)
R.J. Williams, D. Zipser, A learning algorithm for continually running fully recurrent neural networks. Neural Comput. 1, 270–280 (1989)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer Science+Business Media B.V.
About this chapter
Cite this chapter
Endisch, C., Brache, M., Kennel, R. (2010). Parameter Identification of a Nonlinear Two Mass System Using Prior Knowledge. In: Ao, SI., Rieger, B., Amouzegar, M. (eds) Machine Learning and Systems Engineering. Lecture Notes in Electrical Engineering, vol 68. Springer, Dordrecht. https://doi.org/10.1007/978-90-481-9419-3_16
Download citation
DOI: https://doi.org/10.1007/978-90-481-9419-3_16
Published:
Publisher Name: Springer, Dordrecht
Print ISBN: 978-90-481-9418-6
Online ISBN: 978-90-481-9419-3
eBook Packages: EngineeringEngineering (R0)