Time-variant frequency response function measurement in the presence of missing data

Rik Pintelon, Diana Ugryumova, Gerd Vandersteen, Ebrahim Louarroudi, John Lataire

Research output: Unpublished contribution to conferencePoster

Abstract

Time-variant systems are systems with dynamic properties which evolve with time. Examples include: a swinging pendulum with a varying rope length, a civil construction with a varying distribution of the load,... This poster considers that input-output measurements of the time-varying system are available, where some samples of the output are missing. An approach is proposed to estimate the time-varying frequency response function non-parametrically. The problem is formulated as a MISO LTI estimation, with a local polynomial approximation of the linear blocks. This results in a linear least squares problem. It appears that the method is very robust to missing data.
Original languageEnglish
Publication statusPublished - 23 Sep 2018
Event27th ERNSI Workshop on System Identification - Pembroke College, Cambridge, United Kingdom
Duration: 23 Sep 201826 Sep 2018
http://mlg.eng.cam.ac.uk/ernsi2018/

Workshop

Workshop27th ERNSI Workshop on System Identification
Abbreviated titleERNSI 2018
Country/TerritoryUnited Kingdom
CityCambridge
Period23/09/1826/09/18
Internet address

Fingerprint

Dive into the research topics of 'Time-variant frequency response function measurement in the presence of missing data'. Together they form a unique fingerprint.

Cite this