Mechanism for preventing premature convergence in robot localization

Chiang Heng Chien*, Chen Chien Hsu, Wei Yen Wang, Wen Chung Kao

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, we propose a new approach to solve the premature convergence problem in Monte Carlo Localization for global localization, entitled "a mechanism for preventing premature convergence". The algorithm uses a referenced relative vector to rearrange weight for each sample, allowing better exploration of a symmetrical environment, resulting in preventing premature convergence problem to converge particles successfully. Simulations have confirmed that the proposed approach can offer notable improvement in the localization performance.

Original languageEnglish
Title of host publication5th IEEE International Conference on Consumer Electronics - Berlin, ICCE-Berlin 2015
EditorsJose Maria Flores-Arias, Stefan Mozar, Dietmar Hepper, Milan Z. Bjelica, Hans L. Cycon
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages445-446
Number of pages2
ISBN (Electronic)9781479987481
DOIs
Publication statusPublished - 2016 Jan 25
Event5th IEEE International Conference on Consumer Electronics - Berlin, ICCE-Berlin 2015 - Berlin, Germany
Duration: 2015 Sept 62015 Sept 9

Publication series

Name5th IEEE International Conference on Consumer Electronics - Berlin, ICCE-Berlin 2015

Other

Other5th IEEE International Conference on Consumer Electronics - Berlin, ICCE-Berlin 2015
Country/TerritoryGermany
CityBerlin
Period2015/09/062015/09/09

Keywords

  • Global Localization
  • Mobile Robot
  • Monte Carlo Localization
  • Premature Convergence

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Computer Science Applications
  • Hardware and Architecture
  • Media Technology

Fingerprint

Dive into the research topics of 'Mechanism for preventing premature convergence in robot localization'. Together they form a unique fingerprint.

Cite this