Input sequence design for parametric identification of nonlinear systems

We address the issue of designing optimal, plant-friendly input sequences for parametric identification of nonlinear systems. A sequential design methodology based on the constrained optimization of a parameter-space criterion is proposed. The optimization criterion/constraints are chosen to reflect the identification requirements such as information content of the resulting I/O data, control-relevancy, plant-friendly signal characteristics etc. The developments are focused on nonlinear moving average structures. Numerical simulations are presented in order to illustrate the effectiveness of the proposed methodology.