A model error formulation of the multiple model adaptive estimation algorithm

Christopher K. Nebelecky, John L. Crassidis, Puneet Singla

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

4 Scopus citations

Abstract

This paper presents a new form of the multiple model adaptive estimation algorithm for improved state tracking in systems with unknown system models. The proposed approach differs from existing multiple model methods in the manner in which the covariance and Kalman gains of the individual filters are calculated. By using the fused model estimate, recursions for the actual estimation error covariances are derived which account for the deviation of the hypothesized model from the fused model. Using these covariances to determine the Kalman gain leads to improved tracking estimates through fusion of model and measurement uncertainty. The proposed algorithm has been compared against the standard multiple model adaptive estimation and interacting multiple model algorithms in two simulated examples, resulting in improved, and comparable tracking performance, respectively.

Original languageEnglish (US)
Title of host publicationFUSION 2014 - 17th International Conference on Information Fusion
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9788490123553
StatePublished - Oct 3 2014
Event17th International Conference on Information Fusion, FUSION 2014 - Salamanca, Spain
Duration: Jul 7 2014Jul 10 2014

Publication series

NameFUSION 2014 - 17th International Conference on Information Fusion

Other

Other17th International Conference on Information Fusion, FUSION 2014
Country/TerritorySpain
CitySalamanca
Period7/7/147/10/14

All Science Journal Classification (ASJC) codes

  • Information Systems

Fingerprint

Dive into the research topics of 'A model error formulation of the multiple model adaptive estimation algorithm'. Together they form a unique fingerprint.

Cite this