Self-Localization for Mobile Robot Based on Floor Region on Omni-Directional Image.

Accession number;02A0585365
Title;Self-Localization for Mobile Robot Based on Floor Region on Omni-Directional Image.
Author; USUI TOMOYA (Osaka Univ.) FUJIMOTO RYOHEI (Akashi Coll. of Technol.) SEKIMORI DAISUKE (Akashi Coll. of Technol.) MASUTANI YASUHIRO (Osaka Univ.) MIYAZAKI FUMIO (Osaka Univ.)
Journal Title;Nippon Kikai Gakkai Robotikusu, Mekatoronikusu Koenkai Koen Ronbunshu
Journal Code:L0318A
ISSN:
VOL.2001;NO.Pt.3;PAGE.2A1.B3(1)-2A1.B3(2)(2001)
Figure&Table&Reference;FIG.3, TBL.1, REF.3
Pub. Country;Japan
Language;Japanese
Abstract;A method of self-localization based on floor region on omni-directional image is proposed. To overcome occlusion of landmark, a known whole floor region is employed as landmark. After convex hull operation to restore the floor image, the robot position is estimated by non-linear least square method taking account of characteristic of omni-directional image. We confirm effectiveness of the proposed method by experiments using RoboCup small-size league robot equiped with omni-directional vision. (author abst.)