Deakin University
Browse

File(s) not publicly available

Variational Bayesian Kalman filter using natural gradient

journal contribution
posted on 2022-09-30, 00:13 authored by Y HU, X WANG, Q PAN, Z HU, Bill Moran
We propose a technique based on the natural gradient method for variational lower bound maximization for a variational Bayesian Kalman filter. The natural gradient approach is applied to the Kullback-Leibler divergence between the parameterized variational distribution and the posterior density of interest. Using a Gaussian assumption for the parametrized variational distribution, we obtain a closed-form iterative procedure for the Kullback-Leibler divergence minimization, producing estimates of the variational hyper-parameters of state estimation and the associated error covariance. Simulation results in both a Doppler radar tracking scenario and a bearing-only tracking scenario are presented, showing that the proposed natural gradient method outperforms existing methods which are based on other linearization techniques in terms of tracking accuracy.

History

Journal

Chinese Journal of Aeronautics

Volume

35

Issue

5

Pagination

1 - 10

ISSN

1000-9361

Publication classification

C1.1 Refereed article in a scholarly journal

Usage metrics

    Research Publications

    Categories

    No categories selected

    Keywords

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC