A two-tier map representation for compact-stereo-vision-based SLAM

D. C. Herath, S. Kodagoda, G. Dissanayake

Research output: Contribution to journalArticle

Abstract

Vision sensors are increasingly being used in the implementation of Simultaneous Localization and Mapping (SLAM). Even though the mathematical framework of SLAM is well understood, considerable issues remain to be resolved when a particular sensing modality is considered. For instance, the observation model of a small baseline stereo camera is known to be highly nonlinear. As a consequence, state estimations obtained from standard recursive estimators, such as the Extended Kalman Filter, tend to be inconsistent. Further, vision-based approaches are plagued with high feature densities, and the consequent requisite of maintaining large feature databases for loop closure and data association. This paper proposes a two-tier solution for resolving these issues, inspired by the mechanics of human navigation. The proposed two-tier solution addresses the consistency issue by formulating the SLAM problem as a nonlinear batch optimization and presents a novel method for feature management through a two-tier map representation. Simulations and experiments are carried out in an office-like environment to validate the performance of the algorithm.

Original languageEnglish
Pages (from-to)245-256
Number of pages12
JournalRobotica
Volume30
Issue number2
DOIs
Publication statusPublished - Mar 2012
Externally publishedYes

Fingerprint

Simultaneous Localization and Mapping
Stereo vision
Stereo Vision
Data Association
Extended Kalman filters
State Estimation
State estimation
Inconsistent
Kalman Filter
Modality
Batch
Mechanics
Navigation
Baseline
Closure
Sensing
Camera
Cameras
Tend
Estimator

Cite this

Herath, D. C. ; Kodagoda, S. ; Dissanayake, G. / A two-tier map representation for compact-stereo-vision-based SLAM. In: Robotica. 2012 ; Vol. 30, No. 2. pp. 245-256.
@article{e987c7aa5c2d4bfd8b0ce64c0c498bb1,
title = "A two-tier map representation for compact-stereo-vision-based SLAM",
abstract = "Vision sensors are increasingly being used in the implementation of Simultaneous Localization and Mapping (SLAM). Even though the mathematical framework of SLAM is well understood, considerable issues remain to be resolved when a particular sensing modality is considered. For instance, the observation model of a small baseline stereo camera is known to be highly nonlinear. As a consequence, state estimations obtained from standard recursive estimators, such as the Extended Kalman Filter, tend to be inconsistent. Further, vision-based approaches are plagued with high feature densities, and the consequent requisite of maintaining large feature databases for loop closure and data association. This paper proposes a two-tier solution for resolving these issues, inspired by the mechanics of human navigation. The proposed two-tier solution addresses the consistency issue by formulating the SLAM problem as a nonlinear batch optimization and presents a novel method for feature management through a two-tier map representation. Simulations and experiments are carried out in an office-like environment to validate the performance of the algorithm.",
keywords = "Computer vision, Navigation, Pose estimation and registration, Robot localization, SLAM",
author = "Herath, {D. C.} and S. Kodagoda and G. Dissanayake",
year = "2012",
month = "3",
doi = "10.1017/S0263574711000543",
language = "English",
volume = "30",
pages = "245--256",
journal = "Robotica",
issn = "0263-5747",
publisher = "Cambridge University Press",
number = "2",

}

A two-tier map representation for compact-stereo-vision-based SLAM. / Herath, D. C.; Kodagoda, S.; Dissanayake, G.

In: Robotica, Vol. 30, No. 2, 03.2012, p. 245-256.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A two-tier map representation for compact-stereo-vision-based SLAM

AU - Herath, D. C.

AU - Kodagoda, S.

AU - Dissanayake, G.

PY - 2012/3

Y1 - 2012/3

N2 - Vision sensors are increasingly being used in the implementation of Simultaneous Localization and Mapping (SLAM). Even though the mathematical framework of SLAM is well understood, considerable issues remain to be resolved when a particular sensing modality is considered. For instance, the observation model of a small baseline stereo camera is known to be highly nonlinear. As a consequence, state estimations obtained from standard recursive estimators, such as the Extended Kalman Filter, tend to be inconsistent. Further, vision-based approaches are plagued with high feature densities, and the consequent requisite of maintaining large feature databases for loop closure and data association. This paper proposes a two-tier solution for resolving these issues, inspired by the mechanics of human navigation. The proposed two-tier solution addresses the consistency issue by formulating the SLAM problem as a nonlinear batch optimization and presents a novel method for feature management through a two-tier map representation. Simulations and experiments are carried out in an office-like environment to validate the performance of the algorithm.

AB - Vision sensors are increasingly being used in the implementation of Simultaneous Localization and Mapping (SLAM). Even though the mathematical framework of SLAM is well understood, considerable issues remain to be resolved when a particular sensing modality is considered. For instance, the observation model of a small baseline stereo camera is known to be highly nonlinear. As a consequence, state estimations obtained from standard recursive estimators, such as the Extended Kalman Filter, tend to be inconsistent. Further, vision-based approaches are plagued with high feature densities, and the consequent requisite of maintaining large feature databases for loop closure and data association. This paper proposes a two-tier solution for resolving these issues, inspired by the mechanics of human navigation. The proposed two-tier solution addresses the consistency issue by formulating the SLAM problem as a nonlinear batch optimization and presents a novel method for feature management through a two-tier map representation. Simulations and experiments are carried out in an office-like environment to validate the performance of the algorithm.

KW - Computer vision

KW - Navigation

KW - Pose estimation and registration

KW - Robot localization

KW - SLAM

UR - http://www.scopus.com/inward/record.url?scp=84856630330&partnerID=8YFLogxK

U2 - 10.1017/S0263574711000543

DO - 10.1017/S0263574711000543

M3 - Article

VL - 30

SP - 245

EP - 256

JO - Robotica

JF - Robotica

SN - 0263-5747

IS - 2

ER -