GRAY BOX IDENTIFICATION WITH HOPFIELD NEURAL NETWORKS

In this work, a novel method, based upon Hopfield neural networks, is proposed for parameter estimation in the context of system identification. This subject is a very active field of research, because even when a model of a physical system is available, some parameters may be uncertain or time varying. In our methodology, identification is formulated as an optimization problem, profiting from the applicability of Hopfield networks to this kind of problems. In order to compare the novel technique and the classical gradient method, simulations have been carried out for a linearly parameterized system, and results show that the Hopfield network is more efficient than the gradient estimator, obtaining lower error and less oscillations. Also, the neural technique is applied with encouraging results to non-linearly parameterized systems, for which few methods have been proposed.