Identification of Approximative Nonlinear State-Space Models by Subspace
A. Schrempf - Identification of Approximative Nonlinear State-Space Models by Subspace - Proceedings of the 16-th IFAC World Congress, Prag, Tschechische Republik, 2005, pp. 6
A subspace identification algorithm for state-affine state-space systems which allows to approximate nonlinear systems arbitrarily well is derived. The proposed algorithm depends on an approximation step where a detailed approximation error analysis is provided. A special case is presented in which this approximation error vanishes. To tackle higher-order systems and ill-posed problems a regularized kernel method is proposed. The algorithm is evaluated by a simulation study.