Neural identification of linear systems

The authors investigate the use of neural networks for the identification of linear time invariant dynamical systems. Two classes of networks, namely the multilayer feedforward network and the recurrent network with linear neurons are studied. Special attention is devoted to the initialization of weights using prior knowledge of the model structure and parameters, and to a system theory interpretation of neural models. Simulation results enhance the weakness of random initial weights on learning and give some indications for the implementation of the initialization procedures