An integrated weights particle filter algorithm based on correlation particle estimation and sequential importance re-sampling

摘要

In this paper, a new integrated weight particle filter (IWPF) algorithm is proposed based on the combination of correlation particle estimation (CPE) weight and sequential importance re-sampling (SIR) weight. This method can reduce degeneracy phenomenon and re-sampling times of traditional particle filter. By choosing the typical nonlinear system model, the simulation results show that IWPF performs better than CPE and SIR. In our simulation case, this method can provide a 15% increase of accuracy in state estimation and a 30% decrease of re-sampling times.