Variational Bayesian Kalman filter using natural gradient

作     者：Yumei HU Xuezhi WANG Quan PAN Zhentao HU Bill MORAN Yumei HU;Xuezhi WANG;Quan PAN;Zhentao HU;Bill MORAN

作者机构：School of AutomationNorthwestern Polytechnical UniversityXi’an 710072China The Key Laboratory of Information Fusion TechnologyMinistry of EducationXi’an 710072China Department of Electrical and Electronic EngineeringThe University of MelbourneMelbourneVIC 3010Australia School of EngineeringRMIT UniversityMelbourneVIC 3000Australia School of Computer and Information EngineeringHenan UniversityKaifeng 475001China 

出 版 物：《Chinese Journal of Aeronautics》 (中国航空学报（英文版）)

年 卷 期：2022年第35卷第5期

页      面：1-10页

核心收录：

学科分类：080902[工学-电路与系统] 0809[工学-电子科学与技术（可授工学、理学学位）] 08[工学] 

基　　金：co-supported by the National Natural Science Foundation of China(Nos.61790552 and 61976080) the Innovation Foundation for Doctor Dissertation of Northwestern Polytechnical University,China(No.CX201915) 

主　　题：Kullback-Leibler divergence Natural gradient Nonlinear Kalman filter Target tracking Variational Bayesian optimization 

摘      要：We propose a technique based on the natural gradient method for variational lower bound maximization for a variational Bayesian Kalman *** natural gradient approach is applied to the Kullback-Leibler divergence between the parameterized variational distribution and the posterior density of *** 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 *** 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.