Model reduction with pole-zero placement and high order moment matching

Tudor Ionescu*, Orest V. Iftime, Ion Necoara

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

7 Citations (Scopus)
94 Downloads (Pure)

Abstract

In this paper, we calculate a low order model of a linear system of large dimension, that matches a set of high order moments of the transfer function and achieves pole-zero placement constraints. The model satisfying all the constraints simultaneously is selected from a family of parametrized reduced order models. The parameters are computed solving an explicit linear algebraic system. Furthermore, we construct the Loewner matrices from the given data and the imposed pole-zero and first order moment constraints. The resulting approximations achieve a trade-off between good norm approximation and the preservation of the dynamics of the given system in a region of interest. The theory is illustrated on the academic example of the cart controlled by a double pendulum and the practical example of the CD player.

Original languageEnglish
Article number110140
JournalAutomatica
Volume138
DOIs
Publication statusPublished - Apr-2022

Keywords

  • High order moment constraints
  • Loewner matrices
  • Moment matching
  • Pole-zero constraints

Fingerprint

Dive into the research topics of 'Model reduction with pole-zero placement and high order moment matching'. Together they form a unique fingerprint.

Cite this