Experimental investigation of a prediction algorithm for an indoor SLAM platform

Jung Fu Hou, Yu Shin Chou, Yau Zen Chang, Jing Sin Liu*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

This paper presents a scheme for the indoor simultaneous localization and mapping (SLAM) problem. The scheme is based on the scan matching method and is treated as an optimization problem solve by the Simplex method. The two-dimensional distance transform method is used to facilitate the cost value evaluation. In order to register scanned maps with built map through maximum overlap between the maps, a predictive algorithm is proposed. The algorithm can not only reduce search scope but also discard unexpected objects that may cause false match. The approach is investigated by an experimental platform with differential drives. The ICP-SLAM is also implemented for performance comparison. Experimental result shows that the prediction algorithm can improve accumulation error in the indoor environment.

Original languageEnglish
Title of host publicationIntelligent Robotics and Applications - Third International Conference, ICIRA 2010, Proceedings
Pages154-165
Number of pages12
EditionPART 2
DOIs
StatePublished - 2010
Event3rd International Conference on Intelligent Robotics and Applications, ICIRA 2010 - Shanghai, China
Duration: 10 11 201012 11 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume6425 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference3rd International Conference on Intelligent Robotics and Applications, ICIRA 2010
Country/TerritoryChina
CityShanghai
Period10/11/1012/11/10

Keywords

  • Iterative closest point
  • Simplex method
  • Simultaneous localization and mapping

Fingerprint

Dive into the research topics of 'Experimental investigation of a prediction algorithm for an indoor SLAM platform'. Together they form a unique fingerprint.

Cite this