Online Particle Smoothing with Application to Map-matching

12/08/2020

∙

by   Samuel Duffield, et al.

∙

0

∙

share

We introduce a novel method for online smoothing in state-space models based on a fixed-lag approximation. Unlike classical fixed-lag smoothing we approximate the joint posterior distribution rather than just the marginals. By only partially resampling particles, our online particle smoothing technique avoids path degeneracy as the length of the state-space model increases. We demonstrate the utility of our method in the context of map-matching, the task of inferring a vehicle's trajectory given a road network and noisy GPS observations.