PROJECT TITLE :

Particle Smoothing for Conditionally Linear Gaussian Models as Message Passing Over Factor Graphs - 2018

ABSTRACT:

During this Project, the fixed-lag smoothing problem for conditionally linear Gaussian state-space models is investigated from a issue graph perspective. A lot of specifically, once formulating Bayesian smoothing for an arbitrary state-area model as a forward-backward message passing over a issue graph, we tend to target the on top of-mentioned class of models and derive 2 novel particle smoothers for it. Each the proposed techniques are based mostly on the well-known two-filter smoothing approach and employ marginalized particle filtering in their forward pass. But, on the one hand, the first smoothing technique can only be employed to enhance the accuracy of state estimates with respect to that achieved by forward filtering. On the opposite hand, the second technique, that belongs to the class of Rao-Blackwellized particle smoothers, conjointly provides a point mass approximation of the so-known as joint smoothing distribution. Finally, our smoothing algorithms are compared, in terms of estimation accuracy and computational requirements, with a Rao-Blackwellized particle smoother recently proposed by Lindsten et al. (“Rao-Blackwellized particle smoothers for conditionally linear Gaussian models,” IEEE J. Sel. Topics Signal Process., vol. 10, no. two, pp. 353-365, 2016).