Files

EasySLAM_GIT_GVU_10_03.pdf (2.06 MB)

Author(s)

Fathi, Alireza
Cunningham, Alex
Paluri, Balmanohar
Ni, Kai
Dellaert, Frank

Authors

Person

Dellaert, Frank

Advisor(s)

Advisor(s)

Editor(s)

Associated Organization(s)

Organizational Unit

GVU Center

Series

Series

GVU Technical Report Series

Collections

Research Publications

Supplementary to

Permanent Link

http://hdl.handle.net/1853/35126

Abstract

EasySLAM is a robust, accurate, efficient and easy-to-use visual SLAM framework which uses the unique properties of planar landmarks to navigate robots in societal settings. Due to the use of landmarks which can be associated with semantics, a hybrid symbolic-metric SLAM variant is obtained that makes the maps immediately usable for human-robot interaction, high-level monitoring, and semantic analysis. EasySLAM associates a set of landmarks to each part of the house (e.g. kitchen, living room, bathroom, bedroom, etc.) and takes navigation commands such as "go to kitchen". Loalization and mapping, planning and navigation results are presented with an inexpensive, commercially available robot and uniquely identifiable markers. SLAM with planar landmarks is easy, robust, and fills the real need in both research and society, and we have a system that everyone can use.

Sponsor

National Science Foundation

Date Issued

2010

Extent

Resource Type

Text

Resource Subtype

Technical Report

Rights Statement

Rights URI