On The Role of Pre-Filtering in Nonlinear System Identification
| Authors: | Spinelli William, Politecnico di Milano, Italy
Piroddi Luigi, Politecnico di Milano, Italy
Lovera Marco, Politecnico di Milano, Italy |
|---|
| Topic: | 1.1 Modelling, Identification & Signal Processing |
|---|
| Session: | Nonlinear System Identification - Volterra Methods |
|---|
| Keywords: | System identification, Nonlinear systems, Frequency domain, Volterra series, Filtering techniques, Data processing |
|---|
Abstract
Data pre-filtering is often used in linear identification to increase modelaccuracy in a specified frequency band, as pre-filtering is equivalent to afrequency weighting on the prediction error function. However, thisinterpretation applies only to a strictly linear setting of the identificationproblem. In this paper, the role of data and error pre-filtering in nonlinearsystem identification is analyzed and a frequency domain interpretation isprovided, based on the use of Volterra series representations for the truesystem and the considered model structure. Simulation results illustrate theconclusions of the analysis.
