A Variance-Adaptive Particle Filter with Application to Time-Varying Parameter Estimation

Authors:	Zhang Bai, Tsinghua University, China Chen Minze, Tsinghua University, China Zhou D. H., Tsinghua University, China
Topic:	1.4 Stochastic Systems
Session:	Control, Estimation and Analysis of Stochastic Systems
Keywords:	particle filter; sequential Monte Carlo; parameter estimation; variance adaptive

Abstract

Particle filtering, as a new method to solve dynamic system filtering problems, has been applied with great success to many scientific and engineering fields. Particle filters have the ability to perform state estimation in nonlinear and non-Gaussian state space models. However, the standard particle filter algorithm is not applicable for time-varying parameter estimation problems, especially incompetent for abrupt parameters. In this paper, a variance-adaptive particle filter (VAPF) algorithm is proposed, and is applied to time-varying parameter estimation. A simulation example is also presented to demonstrate this method.