Nonparametric identification for the causal optimization of set point tracking

The problem of identifying a nonparametric model of the closed loop part of a 2-d.o.f. control system, suitable for the design of the feedforward regulator, is addressed in this paper. The approach presented to describe the model relies on a finite set of `control´ and `output base functions´, that have to be experimentally determined. A thorough analysis of the methods used to determine these base functions is thus presented, focusing attention on simplicity, feasibility, noise and disturbance sensitivity, as the set point tracking performance of the feedforward regulator depends on the quality of the estimated model.