Resource Efficient Classification of Road Conditions through CNN Pruning

Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review

Autoren

  • Daniel Fink
  • Alexander Busch
  • Mark Wielitzka
  • Tobias Ortmaier

Organisationseinheiten

Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Titel des Sammelwerks21st IFAC World Congress
ErscheinungsortBerlin, Germany
Seiten13958-13963
Seitenumfang6
Band53
Auflage2
PublikationsstatusVeröffentlicht - 2020
Veranstaltung21st IFAC World Congress 2020 - Berlin, Deutschland
Dauer: 12 Juli 2020 → 17 Juli 2020

Publikationsreihe

NameIFAC-PapersOnLine
Herausgeber (Verlag)IFAC Secretariat
ISSN (Print)2405-8963

Abstract

Towards autonomous driving, advanced driver assistance systems increasingly undertake basic driving tasks by replacing human assessment and interactions, when controlling the vehicle. The performance of these systems is directly related to knowledge of the vehicle’s state and influential parameters. In this respect, the road condition has a major influence on the tires’ traction and thus significantly affects the behavior of the vehicle. Therefore, a prediction of the upcoming road condition can improve the performance of the assistance systems which leads to an increased driving safety and comfort. The presented work aims to classify the road surface as well as its weather-related condition, based on images of the front camera view, using deep convolutional neural networks. In order to take computational limitations of vehicle control units into account, a pruning approach is investigated to reduce the network complexity.

ASJC Scopus Sachgebiete

Zitieren

Resource Efficient Classification of Road Conditions through CNN Pruning. / Fink, Daniel; Busch, Alexander; Wielitzka, Mark et al.
21st IFAC World Congress. Band 53 2. Aufl. Berlin, Germany, 2020. S. 13958-13963 (IFAC-PapersOnLine).

Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review

Fink, D, Busch, A, Wielitzka, M & Ortmaier, T 2020, Resource Efficient Classification of Road Conditions through CNN Pruning. in 21st IFAC World Congress. 2 Aufl., Bd. 53, IFAC-PapersOnLine, Berlin, Germany, S. 13958-13963, 21st IFAC World Congress 2020, Berlin, Deutschland, 12 Juli 2020. https://doi.org/10.1016/j.ifacol.2020.12.913
Fink, D., Busch, A., Wielitzka, M., & Ortmaier, T. (2020). Resource Efficient Classification of Road Conditions through CNN Pruning. In 21st IFAC World Congress (2 Aufl., Band 53, S. 13958-13963). (IFAC-PapersOnLine).. https://doi.org/10.1016/j.ifacol.2020.12.913
Fink D, Busch A, Wielitzka M, Ortmaier T. Resource Efficient Classification of Road Conditions through CNN Pruning. in 21st IFAC World Congress. 2 Aufl. Band 53. Berlin, Germany. 2020. S. 13958-13963. (IFAC-PapersOnLine). doi: 10.1016/j.ifacol.2020.12.913
Fink, Daniel ; Busch, Alexander ; Wielitzka, Mark et al. / Resource Efficient Classification of Road Conditions through CNN Pruning. 21st IFAC World Congress. Band 53 2. Aufl. Berlin, Germany, 2020. S. 13958-13963 (IFAC-PapersOnLine).
Download
@inproceedings{63be36b015394c14953fd38584da1819,
title = "Resource Efficient Classification of Road Conditions through CNN Pruning",
abstract = "Towards autonomous driving, advanced driver assistance systems increasingly undertake basic driving tasks by replacing human assessment and interactions, when controlling the vehicle. The performance of these systems is directly related to knowledge of the vehicle{\textquoteright}s state and influential parameters. In this respect, the road condition has a major influence on the tires{\textquoteright} traction and thus significantly affects the behavior of the vehicle. Therefore, a prediction of the upcoming road condition can improve the performance of the assistance systems which leads to an increased driving safety and comfort. The presented work aims to classify the road surface as well as its weather-related condition, based on images of the front camera view, using deep convolutional neural networks. In order to take computational limitations of vehicle control units into account, a pruning approach is investigated to reduce the network complexity.",
keywords = "Classification, Computer vision, Neural networks, Pruning, Road condition",
author = "Daniel Fink and Alexander Busch and Mark Wielitzka and Tobias Ortmaier",
note = "Funding Information: The authors would like to thank the German Research Foundation (DFG) for founding this project.; 21st IFAC World Congress 2020 ; Conference date: 12-07-2020 Through 17-07-2020",
year = "2020",
doi = "10.1016/j.ifacol.2020.12.913",
language = "English",
volume = "53",
series = "IFAC-PapersOnLine",
publisher = "IFAC Secretariat",
pages = "13958--13963",
booktitle = "21st IFAC World Congress",
edition = "2",

}

Download

TY - GEN

T1 - Resource Efficient Classification of Road Conditions through CNN Pruning

AU - Fink, Daniel

AU - Busch, Alexander

AU - Wielitzka, Mark

AU - Ortmaier, Tobias

N1 - Funding Information: The authors would like to thank the German Research Foundation (DFG) for founding this project.

PY - 2020

Y1 - 2020

N2 - Towards autonomous driving, advanced driver assistance systems increasingly undertake basic driving tasks by replacing human assessment and interactions, when controlling the vehicle. The performance of these systems is directly related to knowledge of the vehicle’s state and influential parameters. In this respect, the road condition has a major influence on the tires’ traction and thus significantly affects the behavior of the vehicle. Therefore, a prediction of the upcoming road condition can improve the performance of the assistance systems which leads to an increased driving safety and comfort. The presented work aims to classify the road surface as well as its weather-related condition, based on images of the front camera view, using deep convolutional neural networks. In order to take computational limitations of vehicle control units into account, a pruning approach is investigated to reduce the network complexity.

AB - Towards autonomous driving, advanced driver assistance systems increasingly undertake basic driving tasks by replacing human assessment and interactions, when controlling the vehicle. The performance of these systems is directly related to knowledge of the vehicle’s state and influential parameters. In this respect, the road condition has a major influence on the tires’ traction and thus significantly affects the behavior of the vehicle. Therefore, a prediction of the upcoming road condition can improve the performance of the assistance systems which leads to an increased driving safety and comfort. The presented work aims to classify the road surface as well as its weather-related condition, based on images of the front camera view, using deep convolutional neural networks. In order to take computational limitations of vehicle control units into account, a pruning approach is investigated to reduce the network complexity.

KW - Classification

KW - Computer vision

KW - Neural networks

KW - Pruning

KW - Road condition

UR - http://www.scopus.com/inward/record.url?scp=85105076143&partnerID=8YFLogxK

U2 - 10.1016/j.ifacol.2020.12.913

DO - 10.1016/j.ifacol.2020.12.913

M3 - Conference contribution

VL - 53

T3 - IFAC-PapersOnLine

SP - 13958

EP - 13963

BT - 21st IFAC World Congress

CY - Berlin, Germany

T2 - 21st IFAC World Congress 2020

Y2 - 12 July 2020 through 17 July 2020

ER -