Validation of Underwater Sensor Package Using Feature Based SLAM

Robotic vehicles working in new, unexplored environments must be able to locate themselves in the environment while constructing a picture of the objects in the environment that could act as obstacles that would prevent the vehicles from completing their desired tasks. In enclosed environments, unde...

Full description

Bibliographic Details
Main Authors: Christopher Cain, Alexander Leonessa
Format: Article
Language:English
Published: MDPI AG 2016-03-01
Series:Sensors
Subjects:
Online Access:http://www.mdpi.com/1424-8220/16/3/380
id doaj-2ecc391d4ff14108b57515622dceac15
record_format Article
spelling doaj-2ecc391d4ff14108b57515622dceac152020-11-25T00:46:31ZengMDPI AGSensors1424-82202016-03-0116338010.3390/s16030380s16030380Validation of Underwater Sensor Package Using Feature Based SLAMChristopher Cain0Alexander Leonessa1SEW-EURODRIVE, Lyman, SC 29365, USADepartment of Mechanical Engineering, Center for Dynamic Systems Modeling and Control, Virginia Tech, Blacksburg, VA 24061, USARobotic vehicles working in new, unexplored environments must be able to locate themselves in the environment while constructing a picture of the objects in the environment that could act as obstacles that would prevent the vehicles from completing their desired tasks. In enclosed environments, underwater range sensors based off of acoustics suffer performance issues due to reflections. Additionally, their relatively high cost make them less than ideal for usage on low cost vehicles designed to be used underwater. In this paper we propose a sensor package composed of a downward facing camera, which is used to perform feature tracking based visual odometry, and a custom vision-based two dimensional rangefinder that can be used on low cost underwater unmanned vehicles. In order to examine the performance of this sensor package in a SLAM framework, experimental tests are performed using an unmanned ground vehicle and two feature based SLAM algorithms, the extended Kalman filter based approach and the Rao-Blackwellized, particle filter based approach, to validate the sensor package.http://www.mdpi.com/1424-8220/16/3/380underwater range finderEKF SLAMFastSLAMSLAMvision range findervision odometry
collection DOAJ
language English
format Article
sources DOAJ
author Christopher Cain
Alexander Leonessa
spellingShingle Christopher Cain
Alexander Leonessa
Validation of Underwater Sensor Package Using Feature Based SLAM
Sensors
underwater range finder
EKF SLAM
FastSLAM
SLAM
vision range finder
vision odometry
author_facet Christopher Cain
Alexander Leonessa
author_sort Christopher Cain
title Validation of Underwater Sensor Package Using Feature Based SLAM
title_short Validation of Underwater Sensor Package Using Feature Based SLAM
title_full Validation of Underwater Sensor Package Using Feature Based SLAM
title_fullStr Validation of Underwater Sensor Package Using Feature Based SLAM
title_full_unstemmed Validation of Underwater Sensor Package Using Feature Based SLAM
title_sort validation of underwater sensor package using feature based slam
publisher MDPI AG
series Sensors
issn 1424-8220
publishDate 2016-03-01
description Robotic vehicles working in new, unexplored environments must be able to locate themselves in the environment while constructing a picture of the objects in the environment that could act as obstacles that would prevent the vehicles from completing their desired tasks. In enclosed environments, underwater range sensors based off of acoustics suffer performance issues due to reflections. Additionally, their relatively high cost make them less than ideal for usage on low cost vehicles designed to be used underwater. In this paper we propose a sensor package composed of a downward facing camera, which is used to perform feature tracking based visual odometry, and a custom vision-based two dimensional rangefinder that can be used on low cost underwater unmanned vehicles. In order to examine the performance of this sensor package in a SLAM framework, experimental tests are performed using an unmanned ground vehicle and two feature based SLAM algorithms, the extended Kalman filter based approach and the Rao-Blackwellized, particle filter based approach, to validate the sensor package.
topic underwater range finder
EKF SLAM
FastSLAM
SLAM
vision range finder
vision odometry
url http://www.mdpi.com/1424-8220/16/3/380
work_keys_str_mv AT christophercain validationofunderwatersensorpackageusingfeaturebasedslam
AT alexanderleonessa validationofunderwatersensorpackageusingfeaturebasedslam
_version_ 1725264845367934976