Neural network approach for vision-based track navigation using low-powered computers on MAVs

Khushal Brahmbhatt, Akshatha Rakesh Pai, Sanjay Singh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

A quadrotor Micro Aerial Vehicle (MAV) is designed to navigate a track using neural network approach to identify the direction of the path from a stream of monocular images received from a downward-facing camera mounted on the vehicle. Current autonomous MAVs mainly employ computer vision techniques based on image processing and feature tracking for vision-based navigation tasks. It requires expensive onboard computation and can create latency in the real-time system when working with low-powered computers. By using a supervised image classifier, we shift the costly computational task of training a neural network to classify the direction of the track to an off-board computer. We make use of the learned weights obtained after training to perform simple mathematical operations to predict the class of the image on the onboard computer. We compare the computer vision based tracking approach with the proposed approach to navigate a track using a quadrotor and show that the processing rates of the latter is faster. This allows low-cost, low-powered computers such as the Raspberry Pi to be used efficiently as onboard companion computers for flying vision-based autonomous missions with MAVs.

Original languageEnglish
Title of host publication2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages578-583
Number of pages6
Volume2017-January
ISBN (Electronic)9781509063673
DOIs
Publication statusPublished - 30-11-2017
Event2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017 - Manipal, Mangalore, India
Duration: 13-09-201716-09-2017

Conference

Conference2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017
CountryIndia
CityManipal, Mangalore
Period13-09-1716-09-17

Fingerprint

Micro air vehicle (MAV)
Navigation
Neural networks
Computer vision
Facings
Real time systems
Printed circuit boards
Image processing
Classifiers
Cameras
Antennas
Processing
Costs

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications
  • Computer Science Applications
  • Information Systems

Cite this

Brahmbhatt, K., Pai, A. R., & Singh, S. (2017). Neural network approach for vision-based track navigation using low-powered computers on MAVs. In 2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017 (Vol. 2017-January, pp. 578-583). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICACCI.2017.8125902
Brahmbhatt, Khushal ; Pai, Akshatha Rakesh ; Singh, Sanjay. / Neural network approach for vision-based track navigation using low-powered computers on MAVs. 2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017. Vol. 2017-January Institute of Electrical and Electronics Engineers Inc., 2017. pp. 578-583
@inproceedings{6fd0b1d47ef74cc49dbd279101720c6b,
title = "Neural network approach for vision-based track navigation using low-powered computers on MAVs",
abstract = "A quadrotor Micro Aerial Vehicle (MAV) is designed to navigate a track using neural network approach to identify the direction of the path from a stream of monocular images received from a downward-facing camera mounted on the vehicle. Current autonomous MAVs mainly employ computer vision techniques based on image processing and feature tracking for vision-based navigation tasks. It requires expensive onboard computation and can create latency in the real-time system when working with low-powered computers. By using a supervised image classifier, we shift the costly computational task of training a neural network to classify the direction of the track to an off-board computer. We make use of the learned weights obtained after training to perform simple mathematical operations to predict the class of the image on the onboard computer. We compare the computer vision based tracking approach with the proposed approach to navigate a track using a quadrotor and show that the processing rates of the latter is faster. This allows low-cost, low-powered computers such as the Raspberry Pi to be used efficiently as onboard companion computers for flying vision-based autonomous missions with MAVs.",
author = "Khushal Brahmbhatt and Pai, {Akshatha Rakesh} and Sanjay Singh",
year = "2017",
month = "11",
day = "30",
doi = "10.1109/ICACCI.2017.8125902",
language = "English",
volume = "2017-January",
pages = "578--583",
booktitle = "2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

Brahmbhatt, K, Pai, AR & Singh, S 2017, Neural network approach for vision-based track navigation using low-powered computers on MAVs. in 2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017. vol. 2017-January, Institute of Electrical and Electronics Engineers Inc., pp. 578-583, 2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017, Manipal, Mangalore, India, 13-09-17. https://doi.org/10.1109/ICACCI.2017.8125902

Neural network approach for vision-based track navigation using low-powered computers on MAVs. / Brahmbhatt, Khushal; Pai, Akshatha Rakesh; Singh, Sanjay.

2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017. Vol. 2017-January Institute of Electrical and Electronics Engineers Inc., 2017. p. 578-583.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Neural network approach for vision-based track navigation using low-powered computers on MAVs

AU - Brahmbhatt, Khushal

AU - Pai, Akshatha Rakesh

AU - Singh, Sanjay

PY - 2017/11/30

Y1 - 2017/11/30

N2 - A quadrotor Micro Aerial Vehicle (MAV) is designed to navigate a track using neural network approach to identify the direction of the path from a stream of monocular images received from a downward-facing camera mounted on the vehicle. Current autonomous MAVs mainly employ computer vision techniques based on image processing and feature tracking for vision-based navigation tasks. It requires expensive onboard computation and can create latency in the real-time system when working with low-powered computers. By using a supervised image classifier, we shift the costly computational task of training a neural network to classify the direction of the track to an off-board computer. We make use of the learned weights obtained after training to perform simple mathematical operations to predict the class of the image on the onboard computer. We compare the computer vision based tracking approach with the proposed approach to navigate a track using a quadrotor and show that the processing rates of the latter is faster. This allows low-cost, low-powered computers such as the Raspberry Pi to be used efficiently as onboard companion computers for flying vision-based autonomous missions with MAVs.

AB - A quadrotor Micro Aerial Vehicle (MAV) is designed to navigate a track using neural network approach to identify the direction of the path from a stream of monocular images received from a downward-facing camera mounted on the vehicle. Current autonomous MAVs mainly employ computer vision techniques based on image processing and feature tracking for vision-based navigation tasks. It requires expensive onboard computation and can create latency in the real-time system when working with low-powered computers. By using a supervised image classifier, we shift the costly computational task of training a neural network to classify the direction of the track to an off-board computer. We make use of the learned weights obtained after training to perform simple mathematical operations to predict the class of the image on the onboard computer. We compare the computer vision based tracking approach with the proposed approach to navigate a track using a quadrotor and show that the processing rates of the latter is faster. This allows low-cost, low-powered computers such as the Raspberry Pi to be used efficiently as onboard companion computers for flying vision-based autonomous missions with MAVs.

UR - http://www.scopus.com/inward/record.url?scp=85042915654&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042915654&partnerID=8YFLogxK

U2 - 10.1109/ICACCI.2017.8125902

DO - 10.1109/ICACCI.2017.8125902

M3 - Conference contribution

VL - 2017-January

SP - 578

EP - 583

BT - 2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Brahmbhatt K, Pai AR, Singh S. Neural network approach for vision-based track navigation using low-powered computers on MAVs. In 2017 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2017. Vol. 2017-January. Institute of Electrical and Electronics Engineers Inc. 2017. p. 578-583 https://doi.org/10.1109/ICACCI.2017.8125902