A simple and fast technique to evaluate the possibility of approximating a nonlinear system by a nonlinear PISPO model

In this paper we present a simple approach that allows one to decide whether a nonlinear Periodic Input Same Period Output (PISPO) model can be successfully used for nonlinear system identification, in particular when nonperiodicities are present on the output. The proposed technique consists of a nonparametric preprocessing step based on a periodic analysis, combined with the estimation of the parametric best linear approximation.