Neural Network Based System Identification of an Axis of Car Suspension System

Neural networks system identification have been widely used for estimate the nonlinear model of system. In this paper, multilayer perceptron neural network is used for identifying the Nonlinear AutoRegressive with eXogenous input (NARX) model of a quarter car passive suspension system. Input output data are acquired by driving a car on a special road event. The networks structure is developed based on system model. The Networks learning algorithm is derived using Fisher’s scoring method. Then the Fisher information is given as a weighted covariance matrix of inputs and outputs the network hidden layer. Unitwise Fisher’s scoring method reduces to the algorithm in which each unit estimate its own weights by a weighted least square method. The results show that the method uses suitable for modeling a quarter car passive suspension systems.

[1]  Geoffrey E. Hinton.  Connectionist Learning Procedures , 1989, Artif. Intell..

[2]  John S. Bridle,et al.  Training Stochastic Model Recognition Algorithms as Networks can Lead to Maximum Mutual Information Estimation of Parameters , 1989, NIPS.

[3]  R. Majjad.  Estimation of suspension parameters , 1997, Proceedings of the 1997 IEEE International Conference on Control Applications.

[4]  A. N. Jha,et al.  Nonlinear System Identification using Neural Networks , 2007 .

[5]  Takio Kurita,et al.  Iterative weighted least squares algorithms for neural networks classifiers , 1992, New Generation Computing.

[6]  Geoffrey E. Hinton,et al.  Learning representations by back-propagation errors, nature , 1986 .

[7]  Han-Shue Tan,et al.  Model identification of an automotive hydraulic active suspension system , 1997, Proceedings of the 1997 American Control Conference (Cat. No.97CH36041).

[8]  Hava T. Siegelmann,et al.  Computational capabilities of recurrent NARX neural networks , 1997, IEEE Trans. Syst. Man Cybern. Part B.

[9]  Paul I. Ro,et al.  Reduced-order modelling and parameter estimation for a quarter-car suspension system , 2000 .

[10]  R. Cook.  Generalized Linear Model , 2005 .

[11]  Johan A. K. Suykens,et al.  Nonlinear system identification using neural networks , 1996 .

[12]  P. McCullagh,et al.  Generalized Linear Models , 1984 .

[13]  T D Gillespie,et al.  Fundamentals of Vehicle Dynamics , 1992 .

[14]  Radek Votrubec.  Nonlinear Identification of Global Characteristic of Damping Car System , 2001 .

[15]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.

[16]  G. Hardier.  Recurrent RBF networks for suspension system modeling and wear diagnosis of a damper , 1998, 1998 IEEE International Joint Conference on Neural Networks Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98CH36227).

[17]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[18]  K. T. Schuetze,et al.  INTELLIGENT ESTIMATION OF SYSTEM PARAMETERS FOR ACTIVE VEHICLE SUSPENSION CONTROL , 1999 .