Incremental identification of NARX models by sparse grid approximation
| Authors: | Kahrs Olaf, RWTH Aachen University, Germany
Brendel Marc, RWTH Aachen University, Germany
Marquardt Wolfgang, RWTH Aachen University, Germany |
|---|
| Topic: | 1.1 Modelling, Identification & Signal Processing |
|---|
| Session: | Applications of Nonlinear Modeling Methods |
|---|
| Keywords: | Nonlinear models, sparse grids, input selection, NARX |
|---|
Abstract
Nonlinear empirical models are used in various applications.During model-building, five major steps usually have to be carriedout: model structure selection, determination of input variables,complexity adjustment of the model, parameter estimation and modelvalidation. These steps have to be repeated until a satisfactorymodel is found, which can be very time consuming and may requireuser interaction. This paper proposes an algorithm based on sparsegrid function approximation to incrementally build a nonlinearempirical model. The algorithm exhibits good performance in termsof manual effort and computation time. The method is illustratedby a case study on the identification of a NARX model.
