A biologically inspired improvement strategy for particle filter

Ant colony optimization assisted particle filter

Junpei Zhong, Yu Fai Fung, Mingjun Dai

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

Particle Filter (PF) is a sophisticated model estimation technique based on simulation. Due to the natural limitations of PF, two problems, namely particle impoverishment and sample size dependency, frequently occur during the particles updating stage and these problems will limit the accuracy of the estimation results. In order to alleviate these problems, Ant Colony Optimization is incorporated into the generic PF before the updating stage. After executing the Ant Colony optimization, impoverished particle samples will be re-positioned and closer to their locally highest likelihood distribution function. Our experimental results show that the proposed algorithm can realize better tracking performance when comparing to the generic PF, the Extended Kalman Filter and other enhanced versions of PF.

Original languageEnglish
Pages (from-to)519-526
Number of pages8
JournalInternational Journal of Control, Automation and Systems
Volume8
Issue number3
DOIs
Publication statusPublished - 2010 Jun
Externally publishedYes

Fingerprint

Ant colony optimization
Extended Kalman filters
Distribution functions

Keywords

  • Ant colony optimization
  • Filtering theory
  • Model estimation
  • Particle filters

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Computer Science Applications

Cite this

A biologically inspired improvement strategy for particle filter : Ant colony optimization assisted particle filter. / Zhong, Junpei; Fung, Yu Fai; Dai, Mingjun.

In: International Journal of Control, Automation and Systems, Vol. 8, No. 3, 06.2010, p. 519-526.

Research output: Contribution to journalArticle

@article{09c44c211042485daf9400d4df5d7d1d,
title = "A biologically inspired improvement strategy for particle filter: Ant colony optimization assisted particle filter",
abstract = "Particle Filter (PF) is a sophisticated model estimation technique based on simulation. Due to the natural limitations of PF, two problems, namely particle impoverishment and sample size dependency, frequently occur during the particles updating stage and these problems will limit the accuracy of the estimation results. In order to alleviate these problems, Ant Colony Optimization is incorporated into the generic PF before the updating stage. After executing the Ant Colony optimization, impoverished particle samples will be re-positioned and closer to their locally highest likelihood distribution function. Our experimental results show that the proposed algorithm can realize better tracking performance when comparing to the generic PF, the Extended Kalman Filter and other enhanced versions of PF.",
keywords = "Ant colony optimization, Filtering theory, Model estimation, Particle filters",
author = "Junpei Zhong and Fung, {Yu Fai} and Mingjun Dai",
year = "2010",
month = "6",
doi = "10.1007/s12555-010-0304-7",
language = "English",
volume = "8",
pages = "519--526",
journal = "International Journal of Control, Automation and Systems",
issn = "1598-6446",
publisher = "Institute of Control, Robotics and Systems",
number = "3",

}

TY - JOUR

T1 - A biologically inspired improvement strategy for particle filter

T2 - Ant colony optimization assisted particle filter

AU - Zhong, Junpei

AU - Fung, Yu Fai

AU - Dai, Mingjun

PY - 2010/6

Y1 - 2010/6

N2 - Particle Filter (PF) is a sophisticated model estimation technique based on simulation. Due to the natural limitations of PF, two problems, namely particle impoverishment and sample size dependency, frequently occur during the particles updating stage and these problems will limit the accuracy of the estimation results. In order to alleviate these problems, Ant Colony Optimization is incorporated into the generic PF before the updating stage. After executing the Ant Colony optimization, impoverished particle samples will be re-positioned and closer to their locally highest likelihood distribution function. Our experimental results show that the proposed algorithm can realize better tracking performance when comparing to the generic PF, the Extended Kalman Filter and other enhanced versions of PF.

AB - Particle Filter (PF) is a sophisticated model estimation technique based on simulation. Due to the natural limitations of PF, two problems, namely particle impoverishment and sample size dependency, frequently occur during the particles updating stage and these problems will limit the accuracy of the estimation results. In order to alleviate these problems, Ant Colony Optimization is incorporated into the generic PF before the updating stage. After executing the Ant Colony optimization, impoverished particle samples will be re-positioned and closer to their locally highest likelihood distribution function. Our experimental results show that the proposed algorithm can realize better tracking performance when comparing to the generic PF, the Extended Kalman Filter and other enhanced versions of PF.

KW - Ant colony optimization

KW - Filtering theory

KW - Model estimation

KW - Particle filters

UR - http://www.scopus.com/inward/record.url?scp=77953789713&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77953789713&partnerID=8YFLogxK

U2 - 10.1007/s12555-010-0304-7

DO - 10.1007/s12555-010-0304-7

M3 - Article

VL - 8

SP - 519

EP - 526

JO - International Journal of Control, Automation and Systems

JF - International Journal of Control, Automation and Systems

SN - 1598-6446

IS - 3

ER -