Analysis and comparison of the generic and auxiliary particle filtering frameworks

State estimation is of paramount importance in many fields of engineering. Filtering is the method of estimating the state of a system by incorporating noisy observations as they become available online with prior knowledge of the system model. Particle filters are sequential Monte Carlo methods that use a point mass representation of probability densities in order to propagate the required statistical properties for state estimation. This paper is a quantitative comparison of the generic and auxiliary particle filtering frameworks using various proposal densities and state characterizations. New particle filtering methods that use the extended and unscented Kalman filters as state characterizations in the auxiliary framework are introduced. All the methods are compared in terms of accuracy and. robustness. A synthetic stochastic model that incorporates nonlinear, non-stationary, and non-Gaussian elements is used for the experiments. It is shown that the particle filters designed with the auxiliary framework outperform the generic particle filters and other nonlinear filtering methods in this experiment. This paper is a quantitative comparison of the generic and auxiliary particle filtering frameworks using various proposal densities and state characterizations. New particle filtering methods that use the extended and unscented Kalman filters as state characterizations in the auxiliary framework are introduced. All the methods are compared in terms of accuracy and. robustness. A synthetic stochastic model that incorporates nonlinear, non-stationary, and non-Gaussian elements is used for the experiments. It is shown that the particle filters designed with the auxiliary framework outperform the generic particle filters and other nonlinear filtering methods in this experiment.