Semiparametric identification of Wiener systems using a single harmonic input and retrospective cost optimization

We present a two-step method for identifying SISO Wiener systems. First, using a single harmonic input, we estimate a nonparametric model of the static nonlinearity, which is assumed to be only piecewise continuous. Second, using the identified nonparametric map, we use retrospective cost optimization to identify a parametric model of the linear dynamic system. This method is demonstrated on several examples of increasing complexity.