标题：A unification method of locally identified linear systems and reconstruction of nonlinear dynamical model * * This work was supported in part by JSPS KAKENHI 16H04380.
摘要：AbstractIn this study, we propose an identification method of a state-space model for discrete-time SISO nonlinear systems. The first step of the procedure is to construct models which approximate local dynamics around stationary points of the system using a subspace method. The second step is to describe the local models in one unified coordinate using affine coordinate transformations. The coordinate transformations are determined based on the similarity of adjacent local models by the least-squares method. The model describes quasi-steady behavior of the plant. Simulation results of the proposed method are presented.
关键词：KeywordsIdentificationmodelingSubspace methodsNonlinear system identificationHybriddistributed system identificationIdentification for control
