Federated Kalman filter with the state variables observability degree criterion

Evaluation problem is investigated in the nonlinear formulation. Nonlinear Kalman filter and its implementation methods are presented. Inaccuracy of the evaluated process models adopted in the filter is one of the problems in its implementation. Elimination of this this problem is proposed by using...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhang, Lifei, Neusypin, K. A., Selezneva, M. S., Proletarskiy, A. V., Garina, I. O., Drogovoz, P. A.
Format: Conference Proceeding
Language:English
Subjects:
Criteria
Data integration
Inertial navigation
Kalman filters
Navigation systems
Observability (systems)
Qualitative analysis
State variable
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Evaluation problem is investigated in the nonlinear formulation. Nonlinear Kalman filter and its implementation methods are presented. Inaccuracy of the evaluated process models adopted in the filter is one of the problems in its implementation. Elimination of this this problem is proposed by using the federated Kalman filter. Structural diagrams of the federated Kalman filter are considered. It is proposed to use nonlinear Kalman filters as the local filters. Local filters with models having different quality characteristics are analyzed. Observability degree being a qualitative characteristic was used in selection of models. Determination of the model observability degree is carried out using a numerical criterion of the state variables observability degree in the evaluated process. The observability degree criterion was used to determine correction factors at the information fusion stage in the federated Kalman filter. Effectiveness of the nonlinear federated Kalman filter developed modification is demonstrated by the example of evaluating errors in the aircraft inertial navigation system.
ISSN:0094-243X
1551-7616
DOI:10.1063/5.0111117