Extending filter performance through structured integration

Jarvis Schultz, Todd D. Murphey

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Scopus citations

Abstract

Estimation and filtering form an important component of most modern control systems. Techniques such as extended Kalman filters and particle filters have been successfully utilized for estimation in many different applications. Integrators derived from discrete mechanics possess desirable numerical properties such as stable long-time energy behavior, exact constraint satisfaction, and accurate statistical calculations. In the present work, we leverage these features by utilizing a variational integrator derived from discrete mechanics within extended Kalman filters and particle filters. By filtering real experimental data from the nonlinear, underactuated planar crane problem we demonstrate that the linearizations available through the discrete mechanics framework increase the accuracy of uncertainty estimates provided by an extended Kalman filter, especially when operating at low frequencies. Additionally, we illustrate situations where particle filter performance is increased through the statistics-preserving properties provided by the variational integrator.

Original languageEnglish (US)
Title of host publication2014 American Control Conference, ACC 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages430-436
Number of pages7
ISBN (Print)9781479932726
DOIs
StatePublished - 2014
Event2014 American Control Conference, ACC 2014 - Portland, OR, United States
Duration: Jun 4 2014Jun 6 2014

Publication series

NameProceedings of the American Control Conference
ISSN (Print)0743-1619

Other

Other2014 American Control Conference, ACC 2014
CountryUnited States
CityPortland, OR
Period6/4/146/6/14

Keywords

  • Embedded systems
  • Estimation
  • Modeling and simulation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Extending filter performance through structured integration'. Together they form a unique fingerprint.

Cite this