14th IFAC Symposium on System Identification, SYSID 2006

SYSID-2006 Paper Abstract


Paper ThA2.6

Lacy, Seth (Air Force Res. Lab.), Babuska, Vit (Sandia National Laboratory)

Input-Output Data Scaling for System Identification

Scheduled for presentation during the Regular Session "Identification of Linear Systems II" (ThA2), Thursday, March 30, 2006, 12:10−12:30, Banquet Room

14th IFAC Symposium on System Identification, March 29 - 31, 2006, Newcastle, Australia

This information is tentative and subject to change. Compiled on July 22, 2018

Keywords Identification for Control, Frequency Domain Identification, Multivariable System Identification


System identification, or modeling from data, is used to develop dynamic models for control design, performance prediction, and system analysis. Some system identification algorithms are sensitive to the relative scaling of different input and output channels. In this case, properties of the identified model can vary depending on the scaling of actuators and sensors. Ideally, the identified model would be insensitive to arbitrary scaling of the input and output channels. For system identification algorithms that are scale-sensitive, re-scaling the input and output data can alleviate poor identification results due to initial inappropriate relative scaling of the data. In this paper we describe several methods for choosing weighting matrices for system identification with application to a laboratory experiment.