Modeling Nonlinear Multi-Input Multi-Output Systems using Neural Networks based on Volterra Series

T.I. Haweel and F.A. Alturki (Saudi Arabia)


Modeling and Simulation, Control using NeuralNetworks, Adaptive Control. Non-Linear Control


A new methodology and structure for the neural networks are presented to model nonlinear multi-input multi-output systems. The proposed methodology depends on Volterra series expansion of the input pattern vectors. The proposed structure employs a single layer of neurons with linear transfer functions. This eliminates the hidden layers, the sigmoid non-linear transfer functions and back propagation commonly employed in the conventional neural network (NN). A fast and uniform multi input/output LMS Newton type adaptive algorithm is employed in training the proposed Volterra neural network (VNN) in an incremental mode. The proposed VNN has the great advantage of providing solid and explicit formulas relating the input and target patterns. The improved performance of the VNN over the conventional NN is demonstrated employing a number of simulation experiments.

Important Links:

Go Back