Nonlinear system identification with a feedforward neural network and an optimal bounded ellipsoid algorithm

-Compared to normal learning algorithms, for example backpropagation, the optimal bounded ellipsoid (OBE) algorithm has some better properties, such as faster convergence, since it has a similar structure as the Kalman filter algorithm. Optimal bounded ellipsoid algorithm has some better properties than the Kalman filter training, one is that the noise is not required to be Guassian. In this paper optimal bounded ellipsoid algorithm is applied train the weights of a feedforward neural network for nonlinear system identification. Both hidden layers and output layers can be updated. In order to improve robustness of the optimal of the optimal bounded ellipsoid algorithm, dead-zone is applied to this algorithm. From a dynamic systems point of view, such training can be useful for all neural network applications requiring real-time updating of the weights. Two examples where provided which illustrate the effectiveness of the suggested algorithm based on simulations. Key-Words: Neural Networks, Optimal Bounded Ellipsoid (OBE), Modeling, Identification.