Bayesian Inference for State-Space Models With Student-t Mixture Distributions

This article proposes a robust Bayesian inference approach for linear state-space models with nonstationary and heavy-tailed noise for robust state estimation. The predicted distribution is modeled as the hierarchical Student-<inline-formula> <tex-math notation="LaTeX">$t$ </tex-math></inline-formula> distribution, while the likelihood function is modified to the Student-<inline-formula> <tex-math notation="LaTeX">$t$ </tex-math></inline-formula> mixture distribution. By learning the corresponding parameters online, informative components of the Student-<inline-formula> <tex-math notation="LaTeX">$t$ </tex-math></inline-formula> mixture distribution are adapted to approximate the statistics of potential uncertainties. Then, the obstacle caused by the coupling of the updated parameters is eliminated by the variational Bayesian (VB) technique and fixed-point iterations. Discussions are provided to show the reasons for the achieved advantages analytically. Using the Newtonian tracking example and a three degree-of-freedom (DOF) hover system, we show that the proposed inference approach exhibits better performance compared with the existing method in the presence of modeling uncertainties and measurement outliers.

[1]  Fei Liu,et al.  Sensor Fault Estimation in a Probabilistic Framework for Industrial Processes and its Applications , 2022, IEEE Transactions on Industrial Informatics.

[2]  Fei Liu,et al.  Suboptimal Bayesian state estimators for linear high-dimensional dynamic processes , 2021 .

[3]  Peng Shi,et al.  Variational Adaptive Kalman Filter With Gaussian-Inverse-Wishart Mixture Distribution , 2021, IEEE Transactions on Automatic Control.

[4]  Li Li,et al.  Multi-Party Dynamic State Estimation That Preserves Data and Model Privacy , 2021, IEEE Transactions on Information Forensics and Security.

[5]  Xianqiang Yang,et al.  Nonlinear System Identification With Robust Multiple Model Approach , 2020, IEEE Transactions on Control Systems Technology.

[6]  Shunyi Zhao,et al.  Trial-and-error or avoiding a guess? Initialization of the Kalman filter , 2020, Autom..

[7]  Chunhui Zhao,et al.  Online Probabilistic Estimation of Sensor Faulty Signal in Industrial Processes and Its Applications , 2020, IEEE Transactions on Industrial Electronics.

[8]  Yuriy S. Shmaliy,et al.  Probabilistic Monitoring of Correlated Sensors for Nonlinear Processes in State Space , 2020, IEEE Transactions on Industrial Electronics.

[9]  Taejong Joo,et al.  Being Bayesian about Categorical Probability , 2020, ICML.

[10]  Fang Deng,et al.  A Fast Distributed Variational Bayesian Filtering for Multisensor LTV System With Non-Gaussian Noise , 2019, IEEE Transactions on Cybernetics.

[11]  Qing-Long Han,et al.  A Threshold-Parameter-Dependent Approach to Designing Distributed Event-Triggered $H_{\infty}$ Consensus Filters Over Sensor Networks , 2019, IEEE Transactions on Cybernetics.

[12]  Yanjun Ma,et al.  Probabilistic Monitoring of Sensors in State-Space With Variational Bayesian Inference , 2019, IEEE Transactions on Industrial Electronics.

[13]  Henry Leung,et al.  Student-t mixture labeled multi-Bernoulli filter for multi-target tracking with heavy-tailed noise , 2018, Signal Process..

[14]  Yonggang Zhang,et al.  A Novel Adaptive Kalman Filter With Inaccurate Process and Measurement Noise Covariance Matrices , 2018, IEEE Transactions on Automatic Control.

[15]  Yonggang Zhang,et al.  A Novel Robust Student's t-Based Kalman Filter , 2017, IEEE Transactions on Aerospace and Electronic Systems.

[16]  Yonggang Zhang,et al.  Robust student’s t based nonlinear filter and smoother , 2016, IEEE Transactions on Aerospace and Electronic Systems.

[17]  Roland Hostettler,et al.  Sigma-point filtering for nonlinear systems with non-additive heavy-tailed noise , 2016, 2016 19th International Conference on Information Fusion (FUSION).

[18]  Yonggang Zhang,et al.  A robust Gaussian approximate filter for nonlinear systems with heavy tailed measurement noises , 2016, 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[19]  Fredrik Gustafsson,et al.  A Student's t filter for heavy tailed process and measurement noise , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[20]  Henry Leung,et al.  A variational Bayesian approach to robust sensor fusion based on Student-t distribution , 2013, Inf. Sci..

[21]  Eduardo Mario Nebot,et al.  Approximate Inference in State-Space Models With Heavy-Tailed Noise , 2012, IEEE Transactions on Signal Processing.

[22]  Lamine Mili,et al.  Robust Kalman Filter Based on a Generalized Maximum-Likelihood-Type Estimator , 2010, IEEE Transactions on Signal Processing.

[23]  Simo Särkkä,et al.  Recursive Noise Adaptive Kalman Filtering by Variational Bayesian Approximations , 2009, IEEE Transactions on Automatic Control.

[24]  Tom Drummond,et al.  Student-tMixture Filter for Robust, Real-Time Visual Tracking , 2008, ECCV.

[25]  Shie Mannor,et al.  A Kalman Filter Design Based on the Performance/Robustness Tradeoff , 2008, IEEE Transactions on Automatic Control.

[26]  Pedro Albertos,et al.  A new dead-time compensator to control stable and integrating processes with long dead-time , 2008, Autom..

[27]  James B. Rawlings,et al.  A new autocovariance least-squares method for estimating noise covariances , 2006, Autom..

[28]  Raman K. Mehra,et al.  Approaches to adaptive filtering , 1970 .

[29]  R. Kashyap.  Maximum likelihood identification of stochastic linear systems , 1970 .

[30]  Rosen D Von.  On moments of the inverted Wishart distribution , 1997 .

[31]  B. Tapley,et al.  Adaptive sequential estimation with unknown noise statistics , 1976 .