Design of experiments for nonlinear system identification: a set membership approach (Q2207177)

Data-driven nonlinear dynamic systems identification is important generic approach for building mathematical models of dynamic systems. It origins from the classic Weierstrass approximation theorem and Frénchet's results that a continuous real functional defined on a compact set of real continuous functions could be approximated by the sum of a finite number of the Volterra functional series' terms. Here readers may refer to the monograph by \textit{D. Sidorov} [Integral dynamical models. Singularities, signals and control. Hackensack, NJ: World Scientific (2015; Zbl 1311.45012)] and its bibliography. The first step here is constraction of the training/identification dataset introduced under the terms ``design of experiment (DoE)''. The contribution of this paper is threefold. First, quasi-local nonlinear set membership (SM) identification method is presented. Second, the set membership DoE algorithm is proposed. It minimizes the worst-case model error. Finally, online DoE algorithm is described and validated on two simulation examples. Results of this paper can be employed to design an adaptive data-driven nonlinear dynamic system control.