Map Building and Terrain-Aided Localisation in an Underground Mine - Robotics Institute Carnegie Mellon University

Map Building and Terrain-Aided Localisation in an Underground Mine

G. Dissanayake, Martial Hebert, Anthony (Tony) Stentz, and H. Durrant-Whyte
Conference Paper, Proceedings of 1st International Conference on Field and Service Robotics (FSR '97), pp. 40 - 45, December, 1997

Abstract

This paper discusses strategies for estimation of the location of a Load-Haul-Dump truck (LHD) in and underground mine. The results reported are based on data obtained from field trials in an underground mine in Queensland. An LHD was fitted with two two-dimensional laser rangefinders, an INS, sensors to measure vehicle speed and the articulation angle and a rotating laser scanner. An algorithm based on an extended Kaiman filter provides a poly-line map of the mine walls. Once the map is built, the iterative closest point (ICP) algorithm is used to obtain an estimate of the location of the LHD using successive scans from the laser rangefinder.

BibTeX

@conference{Dissanayake-1997-16398,
author = {G. Dissanayake and Martial Hebert and Anthony (Tony) Stentz and H. Durrant-Whyte},
title = {Map Building and Terrain-Aided Localisation in an Underground Mine},
booktitle = {Proceedings of 1st International Conference on Field and Service Robotics (FSR '97)},
year = {1997},
month = {December},
pages = {40 - 45},
}