Abstract

Navigation is one of the most complex daily activities we engage in. Partly due to its complexity, navigational abilities are vulnerable to many conditions including Topographical Agnosia, Alzheimer’s Disease, and vision impairments. While navigation using solely vision remains a difficult problem in the field of assistive technology, emerging methods in Deep Reinforcement Learning and Computer Vision show promise in producing vision-based navigational aids for those with navigation impairments. To this effect, we introduce GraphMem, a Neural Computing approach to navigation tasks and compare it to several state of the art Neural Computing methods in a one-shot, 3D, first-person maze solving task. Comparing GraphMem to current methods in navigation tasks unveils insights into navigation and represents a first step towards employing these emerging techniques in navigational assistive technology.

Original languageEnglish (US)
Title of host publicationSmart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers
EditorsStefano Berretti, Anup Basu
PublisherSpringer Verlag
Pages66-75
Number of pages10
ISBN (Print)9783030043742
DOIs
StatePublished - Jan 1 2018
Event1st International Conference on Smart Multimedia, ICSM 2018 - Toulon, France
Duration: Aug 24 2018Aug 26 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11010 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other1st International Conference on Smart Multimedia, ICSM 2018
CountryFrance
CityToulon
Period8/24/188/26/18

Fingerprint

Reinforcement learning
Reinforcement Learning
Navigation
Assistive Technology
Computing Methods
Alzheimer's Disease
Computer Vision
Computer vision
Person
Computing
Vision

Keywords

  • Assistive technology
  • Navigation
  • Reinforcement learning
  • Topographical agnosia

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Fakhri, B., Keech, A., Schlosser, J., Brooks, E., Demakethepalli Venkateswara, H., Panchanathan, S., & Kira, Z. (2018). Deep reinforcement learning methods for navigational aids. In S. Berretti, & A. Basu (Eds.), Smart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers (pp. 66-75). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11010 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-04375-9_6

Deep reinforcement learning methods for navigational aids. / Fakhri, Bijan; Keech, Aaron; Schlosser, Joel; Brooks, Ethan; Demakethepalli Venkateswara, Hemanth; Panchanathan, Sethuraman; Kira, Zsolt.

Smart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers. ed. / Stefano Berretti; Anup Basu. Springer Verlag, 2018. p. 66-75 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11010 LNCS).

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

Fakhri, B, Keech, A, Schlosser, J, Brooks, E, Demakethepalli Venkateswara, H, Panchanathan, S & Kira, Z 2018, Deep reinforcement learning methods for navigational aids. in S Berretti & A Basu (eds), Smart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11010 LNCS, Springer Verlag, pp. 66-75, 1st International Conference on Smart Multimedia, ICSM 2018, Toulon, France, 8/24/18. https://doi.org/10.1007/978-3-030-04375-9_6
Fakhri B, Keech A, Schlosser J, Brooks E, Demakethepalli Venkateswara H, Panchanathan S et al. Deep reinforcement learning methods for navigational aids. In Berretti S, Basu A, editors, Smart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers. Springer Verlag. 2018. p. 66-75. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-04375-9_6
Fakhri, Bijan ; Keech, Aaron ; Schlosser, Joel ; Brooks, Ethan ; Demakethepalli Venkateswara, Hemanth ; Panchanathan, Sethuraman ; Kira, Zsolt. / Deep reinforcement learning methods for navigational aids. Smart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers. editor / Stefano Berretti ; Anup Basu. Springer Verlag, 2018. pp. 66-75 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d3c6a80d9eaf4c0482a09ed8a7d007e9,
title = "Deep reinforcement learning methods for navigational aids",
abstract = "Navigation is one of the most complex daily activities we engage in. Partly due to its complexity, navigational abilities are vulnerable to many conditions including Topographical Agnosia, Alzheimer’s Disease, and vision impairments. While navigation using solely vision remains a difficult problem in the field of assistive technology, emerging methods in Deep Reinforcement Learning and Computer Vision show promise in producing vision-based navigational aids for those with navigation impairments. To this effect, we introduce GraphMem, a Neural Computing approach to navigation tasks and compare it to several state of the art Neural Computing methods in a one-shot, 3D, first-person maze solving task. Comparing GraphMem to current methods in navigation tasks unveils insights into navigation and represents a first step towards employing these emerging techniques in navigational assistive technology.",
keywords = "Assistive technology, Navigation, Reinforcement learning, Topographical agnosia",
author = "Bijan Fakhri and Aaron Keech and Joel Schlosser and Ethan Brooks and {Demakethepalli Venkateswara}, Hemanth and Sethuraman Panchanathan and Zsolt Kira",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-030-04375-9_6",
language = "English (US)",
isbn = "9783030043742",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "66--75",
editor = "Stefano Berretti and Anup Basu",
booktitle = "Smart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers",

}

TY - GEN

T1 - Deep reinforcement learning methods for navigational aids

AU - Fakhri, Bijan

AU - Keech, Aaron

AU - Schlosser, Joel

AU - Brooks, Ethan

AU - Demakethepalli Venkateswara, Hemanth

AU - Panchanathan, Sethuraman

AU - Kira, Zsolt

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Navigation is one of the most complex daily activities we engage in. Partly due to its complexity, navigational abilities are vulnerable to many conditions including Topographical Agnosia, Alzheimer’s Disease, and vision impairments. While navigation using solely vision remains a difficult problem in the field of assistive technology, emerging methods in Deep Reinforcement Learning and Computer Vision show promise in producing vision-based navigational aids for those with navigation impairments. To this effect, we introduce GraphMem, a Neural Computing approach to navigation tasks and compare it to several state of the art Neural Computing methods in a one-shot, 3D, first-person maze solving task. Comparing GraphMem to current methods in navigation tasks unveils insights into navigation and represents a first step towards employing these emerging techniques in navigational assistive technology.

AB - Navigation is one of the most complex daily activities we engage in. Partly due to its complexity, navigational abilities are vulnerable to many conditions including Topographical Agnosia, Alzheimer’s Disease, and vision impairments. While navigation using solely vision remains a difficult problem in the field of assistive technology, emerging methods in Deep Reinforcement Learning and Computer Vision show promise in producing vision-based navigational aids for those with navigation impairments. To this effect, we introduce GraphMem, a Neural Computing approach to navigation tasks and compare it to several state of the art Neural Computing methods in a one-shot, 3D, first-person maze solving task. Comparing GraphMem to current methods in navigation tasks unveils insights into navigation and represents a first step towards employing these emerging techniques in navigational assistive technology.

KW - Assistive technology

KW - Navigation

KW - Reinforcement learning

KW - Topographical agnosia

UR - http://www.scopus.com/inward/record.url?scp=85058549481&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058549481&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-04375-9_6

DO - 10.1007/978-3-030-04375-9_6

M3 - Conference contribution

AN - SCOPUS:85058549481

SN - 9783030043742

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 66

EP - 75

BT - Smart Multimedia - 1st International Conference, ICSM 2018, Revised Selected Papers

A2 - Berretti, Stefano

A2 - Basu, Anup

PB - Springer Verlag

ER -