TY - GEN
T1 - Variational compensation based nonlinear filter for continuous-discrete stochastic systems
AU - Wang, Tingjun
AU - Cui, Haoran
AU - Wang, Xiaoxu
N1 - Publisher Copyright:
© 2020 International Society of Information Fusion (ISIF).
PY - 2020/7
Y1 - 2020/7
N2 - In this paper, a novel variational compensation based nonlinear filter (VCNF) is proposed to cope with the nonlinear filtering problem in continuous-discrete systems. The core of VCNF is to construct a variational state compensation model with variational compensation parameters for accurately describing uncertain continuous state. The role of variational compensation parameters is to adaptively compensate the unpredictable approximation and discretization errors of system states. In the variational Bayesian framework, through iteratively and alternatively achieving the fitting of the state priori model and the compensation of approximation and discretization errors, estimation accuracy and adaptiveness can be enhanced gradually. The superior performance of VCNF is demonstrated in the simulation of target tracking.
AB - In this paper, a novel variational compensation based nonlinear filter (VCNF) is proposed to cope with the nonlinear filtering problem in continuous-discrete systems. The core of VCNF is to construct a variational state compensation model with variational compensation parameters for accurately describing uncertain continuous state. The role of variational compensation parameters is to adaptively compensate the unpredictable approximation and discretization errors of system states. In the variational Bayesian framework, through iteratively and alternatively achieving the fitting of the state priori model and the compensation of approximation and discretization errors, estimation accuracy and adaptiveness can be enhanced gradually. The superior performance of VCNF is demonstrated in the simulation of target tracking.
KW - Continuous-discrete stochastic system
KW - Nonlinear Kalman filter
KW - Target tracking
KW - Variational Bayesian method
UR - http://www.scopus.com/inward/record.url?scp=85092740170&partnerID=8YFLogxK
U2 - 10.23919/FUSION45008.2020.9190435
DO - 10.23919/FUSION45008.2020.9190435
M3 - 会议稿件
AN - SCOPUS:85092740170
T3 - Proceedings of 2020 23rd International Conference on Information Fusion, FUSION 2020
BT - Proceedings of 2020 23rd International Conference on Information Fusion, FUSION 2020
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 23rd International Conference on Information Fusion, FUSION 2020
Y2 - 6 July 2020 through 9 July 2020
ER -