Identification of nonlinear dynamical systems using multilayered neural networks (Q675277)

The paper deals with the open-loop identification of four classes of discrete-time, multi-input multi-output, nonlinear systems. The structure of the neural-net identifier is derived using a passivity approach, and the weight tuning algorithm relies on a slightly modified delta rule. It is shown that the persistency-of-excitation condition guarantees the convergence of the identification error and the boundedness of all the signals in the system. The certainty equivalence assumption is not used.