Particle Filtering Combined with Interval Methods for Tracking Applications

Abstract

This chapter introduces a new approach for sequential nonlinear estimation based on a combination of particle filtering and interval analysis for tracking applications. The method is first presented within the interval framework, by introducing a new concept of box particles for the purposes of drastically reducing the number of particles. In this chapter, the box particle filter (Box-PF) is presented in its original ad hoc formulation. The chapter provides an overview of the Bayesian inference methodology. The chapter gives a theoretical derivation of the Box-PF as a sum of uniform probability density functions (pdfs). It demonstrates the advantages of the Box-PF over a dynamic localization example. The contributions and open issues for future works are summarized in the chapter.