A Federated particle filtering algorithm based on EKPF

Author

Meng, Li ; Jia-hong, Liang

Author_Institution

Coll. of Mech. Eng. & Autom., Nat. Univ. of Defence Technol., Changsha, China

fYear

2011

fDate

15-17 April 2011

Firstpage

2873

Lastpage

2876

Abstract

To solve the problem of information fusion in the system of nonlinear/non-Gaussian error models, a new algorithm called federated EKPF (FEKPF) algorithm was proposed based on EKPF and federated Kalman filter. In this algorithm, EKPF served as the local filter, whereas the master filter adopt the information fusion algorithm to obtain the global state estimations which are transmitted to each local filter to update particles as the feedbacks according to information distribution. The proposed algorithm was tested in subsequent simulation by contrast to FEKF. The results showed that FEKPF is more effective for nonlinear/non-Gaussian systems.

Keywords

Kalman filters; nonlinear filters; particle filtering (numerical methods); sensor fusion; state estimation; EKPF; extended Kalman particle filter; federated Kalman filter; federated particle filtering algorithm; global state estimations; information distribution; information fusion problem; master filter; nonGaussian error models; nonlinear error models; Bayesian methods; Filtering algorithms; Global Positioning System; Information filters; Kalman filters; Particle filters; extended Kalman particle filter; federated extended Kalman filter; information fusion;

fLanguage

English

Publisher

ieee

Conference_Titel

Electric Information and Control Engineering (ICEICE), 2011 International Conference on

Conference_Location

Wuhan

Print_ISBN

978-1-4244-8036-4

Type

conf

DOI

10.1109/ICEICE.2011.5778115

Filename

5778115