Identification structures using rational wavelets: Examples of application. (Q1417613)

The authors propose an identification structure consisting of two blocks in cascade: the first one uses rational wavelets for identifying the linear dynamic part, while for the nonlinear static part they choose a single hidden layer perception neural network, or a basis of high level canonical piecewise linear function. Thus, this structure is the traditional Wiener model, which may be viewed as a black-box model with semi-physical regressors, and allows one to approximate the dynamic evolution of any nonlinear, causal, time-invariant system with fading memory. In order to illustrate the application of the above black-box identification structure, they present three examples, where the two examples use data sets from numerical simulations and the third example uses a data set from an actuated pendulum experiment.