Particle filter based on PSO

The main challenge in using particle filter (PF) to nonlinear state estimation problem is the particle degeneracy. Resampling operation solves degeneracy to some extent, but it results in the phenomenon of sample impoverishment. Therefore, it cannot achieve the satisfactory accuracy generally with certain number particles by using generic PF algorithm because of the serious impoverishment problem. Here we aim for decreasing the impoverishment of samples set after resampling step. The principle of PF together with its particle degeneracy and sample impoverishment problems are introduced in this paper. Based on the analysis of the causes of sample impoverishment, particle swarm optimization (PSO) which is one of the swarm intelligence algorithms is introduced to PF to ameliorate the diversity of samples set after resampling step. Thus a new algorithm which is called PSO-PF is proposed. From a theoretical analysis, the PSO operation on particles set can overcome sample impoverishment problem largely. And finally, a generic numerical example shows that PSO-PF presents better than generic PF algorithm regarding to accuracy.