Three Dimensional Evidence Grids for SLAM in Complex Underwater Environments - Robotics Institute Carnegie Mellon University

Three Dimensional Evidence Grids for SLAM in Complex Underwater Environments

Conference Paper, Proceedings of 14th International Symposium of Unmanned Untethered Submersible Technology (UUST '05), August, 2005

Abstract

We describe the initial results in constructing the simultaneous localization and mapping system for exploration of flooded subterranean environments like mines and caves for an autonomous hovering vehicle which will explore flooded cenotes (sinkholes) in Mexico. In May 2005, a preliminary data set was collected in the Zacaton cenote using a drop sonde equipped with a circular array of 32 pencil beam sonar transducers. This paper presents the results of applying an evidence grid and particle filter localization system to the Zacaton data, as well as some simulations that investigate the effectiveness of different sonar geometry configurations.

BibTeX

@conference{Fairfield-2005-9254,
author = {Nathaniel Fairfield and George A. Kantor and David Wettergreen},
title = {Three Dimensional Evidence Grids for SLAM in Complex Underwater Environments},
booktitle = {Proceedings of 14th International Symposium of Unmanned Untethered Submersible Technology (UUST '05)},
year = {2005},
month = {August},
publisher = {AUSI},
address = {Lee, New Hampshire},
keywords = {evidence grids, slam, underwater},
}