Bayesian estimation for target tracking: part II, the Gaussian sigma‐point Kalman filters

Abstract

This is the second part of a three part article examining methods for Bayesian estimation and tracking. In the first part we presented the general theory of Bayesian estimation where we showed that Bayesian estimation methods can be divided into two very general classes: a class where the observation conditioned posterior densities are propagated in time through a predictor/corrector method; and a second class where the first two moments are propagated in time, with state and observation moment prediction steps followed by state moment update steps that use the latest observations. In this second part, we make the assumption that all densities are Gaussian and, after applying an affine transformation and approximating all nonlinear functions by interpolating polynomials, we recover the sigma‐point class of Kalman filters, including the unscented, spherical simplex, and Gauss‐Hermite Kalman filters. In part 3, we will show that approximating a density by a set of Monte Carlo samples leads to particle filter methods, where the posterior density is propagated in time and moment integrals are approximated by sample moments. WIREs Comput Stat 2012 doi: 10.1002/wics.1215 This article is categorized under: Statistical and Graphical Methods of Data Analysis > Bayesian Methods and Theory

On the ubiquity of the Bayesian paradigm in statistical machine learning and data science

“…We see the power of the Bayesian thought directly or indirectly in state of the art settings such Latent Dirichlet Allocation (LDA) for Topic Modelling [3]. The forever useful Kalman filter, thanks to the latent space has also benefitted heavily from the power of the Bayesian paradigm [20,21] and [22]. Even before the blessings of affordable computation ushered in the glorious era of the Bayesian thought, Markov Random Fields were being used for the Statistical Analysis of Dirty Pictures [2], already anchoring the palpable power of the Gospel according to Reverend Thomas Bayes.…”

Bayesian estimation for target tracking, Part III: Monte Carlo filters

“…In Part II2 of the three part article series, we examined the methods for estimating the solution for Eq. (2) by approximating \documentclass{article}\usepackage{amsmath}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{amsfonts}\pagestyle{empty}\begin{document}$\mathbf{\widetilde{g}}$ \end{document}( c ) by some form of polynomial.…”

“…(4) becomes We find that Eq. (7) has the exact form needed to utilize the sigma point process shown in Figure 4 of Ref 2, resulting in the sigma point MCKF.…”

“…In the second part,2 we examined the specific case for propagating moments when all densities are Gaussian and all nonlinear functions are approximated by general interpolating polynomials. Approximation of the nonlinear functions by general interpolating polynomials resulted in predictive integrals that are Gaussian‐weighted integrals over the multidimensional monomials associated with each term in the polynomial.…”

Multi-Rate Data Fusion for State and Parameter Estimation in (Bio-)Chemical Process Engineering

“…In [3], the most significant estimation algorithms have been presented and compared. Here, Bayesian estimators represent a specific category [4][5][6][7][8] that is an alternative to optimization based techniques such as moving horizon estimators [9][10][11]. An important class of approximative Bayesian estimators are Kalman filters (KF) and particle filters [6,12,13].…”