Resampling Methods for Particle Filtering: Classification, implementation, and strategies

TLDR

Particle filtering has resurged over the past two decades, gaining popularity for processing nonlinear state‑space models with non‑Gaussian noise and being applied across finance, geophysics, communications, control, navigation, tracking, and robotics, prompting numerous review articles. The study aims to supply practitioners and researchers with guidelines for selecting and applying resampling techniques in particle filtering. The authors classify resampling methods and compare their properties within a proposed framework, focusing on qualitative descriptions of the algorithms. The article reviews the state of the art of resampling methods for particle filtering.

Abstract

Two decades ago, with the publication, we witnessed the rebirth of particle filtering (PF) as a methodology for sequential signal processing. Since then, PF has become very popular because of its ability to process observations represented by nonlinear state-space models where the noises of the model can be non-Gaussian. This methodology has been adopted in various fields, including finance, geophysical systems, wireless communications, control, navigation and tracking, and robotics. The popularity of PF has also spurred the publication of several review articles. In this article, the state of the art of resampling methods was reviewed. The methods were classified and their properties were compared in the framework of the proposed classifications. The emphasis in the article was on the classification and qualitative descriptions of the algorithms. The intention was to provide guidelines to practitioners and researchers.