Estimation of States of Nonlinear Systems using a Particle Filter

Abstract

Particle filters can estimate the states of nonlinear and non-Gaussian systems without any approximation when the number of particles tends to infinity. However, the method is not popular in industry because the implementation details are missing in the literature. In this paper we discuss several implementation issues and propose novel techniques for tuning the particle filter and dealing with multi-rate data. The performance of the proposed methodologies are demonstrated using a simulated non-linear CSTR and an experimental four tank system.