Machine learning algorithms for wet road surface detection using acoustic measurements

Stratis Kanarachos; Mike Blundell; Olivier Haas; Michalis Kalliris; Rigas Kotsakis

doi:10.1109/ICMECH.2019.8722834

Machine learning algorithms for wet road surface detection using acoustic measurements

Stratis Kanarachos
, Mike Blundell
, Olivier Haas
, Michalis Kalliris
, Rigas Kotsakis

Aristotle University of Thessaloniki

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

40 Citations (Scopus)

Abstract

—Precipitation can adversely influence road safety.
Slippery road conditions have traditionally been detected using
reactive methods requiring considerable excitation of the tire
forces. Alternatives rely on non-contact methods such as vision,
sound or ultrasonic sensors. This study proposes a cost-effective
wet road conditions detection method based on acoustic
measurements for urban and highway driving. It compared the
performance of a range of machine learning algorithms to
classify the road condition based on the audio features
calculated using octave-band frequency analysis. The approach
was evaluated experimentally using data collected from a
vehicle instrumented with a microphone, GPS and CAN bus
data logger. Support Vector Machines using Quadratic and
Cubic kernels, as well as Logistic Regression performed better
compared to other machine learning-based methods

Original language	English
Title of host publication	Proceedings - 2019 IEEE International Conference on Mechatronics, ICM 2019
Publisher	IEEE
Pages	265-270
Number of pages	6
Volume	(In-press)
ISBN (Electronic)	9781538669594
DOIs	https://doi.org/10.1109/ICMECH.2019.8722834
Publication status	Published - 24 May 2019
Event	IEEE 2019 International Conference on Mechatronics - Technische Universität Ilmenau, Ilmenau, Germany Duration: 19 Mar 2019 → 21 Mar 2019 https://ieee-icm2019.org

Conference

Conference	IEEE 2019 International Conference on Mechatronics
Country/Territory	Germany
City	Ilmenau
Period	19/03/19 → 21/03/19
Internet address	https://ieee-icm2019.org

Keywords

acoustic measurements
wet road surface detection

ASJC Scopus subject areas

Artificial Intelligence
Human-Computer Interaction
Automotive Engineering
Mechanical Engineering
Control and Optimization
Industrial and Manufacturing Engineering

Access to Document

10.1109/ICMECH.2019.8722834

Cite this

Machine learning algorithms for wet road surface detection using acoustic measurements. / Kanarachos, Stratis; Blundell, Mike ; Haas, Olivier et al.
Proceedings - 2019 IEEE International Conference on Mechatronics, ICM 2019. Vol. (In-press) IEEE, 2019. p. 265-270 8722834.

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

Kanarachos, S, Blundell, M , Haas, O, Kalliris, M & Kotsakis, R 2019, Machine learning algorithms for wet road surface detection using acoustic measurements. in Proceedings - 2019 IEEE International Conference on Mechatronics, ICM 2019. vol. (In-press), 8722834, IEEE, pp. 265-270, IEEE 2019 International Conference on Mechatronics, Ilmenau, Thuringia, Germany, 19/03/19. https://doi.org/10.1109/ICMECH.2019.8722834

@inproceedings{5c5c378eb69845cea6335b2a0e5e67cc,

title = "Machine learning algorithms for wet road surface detection using acoustic measurements",

abstract = "—Precipitation can adversely influence road safety. Slippery road conditions have traditionally been detected using reactive methods requiring considerable excitation of the tire forces. Alternatives rely on non-contact methods such as vision, sound or ultrasonic sensors. This study proposes a cost-effective wet road conditions detection method based on acoustic measurements for urban and highway driving. It compared the performance of a range of machine learning algorithms to classify the road condition based on the audio features calculated using octave-band frequency analysis. The approach was evaluated experimentally using data collected from a vehicle instrumented with a microphone, GPS and CAN bus data logger. Support Vector Machines using Quadratic and Cubic kernels, as well as Logistic Regression performed better compared to other machine learning-based methods",

keywords = "acoustic measurements, wet road surface detection",

author = "Stratis Kanarachos and Mike Blundell and Olivier Haas and Michalis Kalliris and Rigas Kotsakis",

year = "2019",

month = may,

day = "24",

doi = "10.1109/ICMECH.2019.8722834",

language = "English",

volume = "(In-press)",

pages = "265--270",

booktitle = "Proceedings - 2019 IEEE International Conference on Mechatronics, ICM 2019",

publisher = "IEEE",

address = "United States",

note = "IEEE 2019 International Conference on Mechatronics ; Conference date: 19-03-2019 Through 21-03-2019",

url = "https://ieee-icm2019.org",

}

TY - GEN

T1 - Machine learning algorithms for wet road surface detection using acoustic measurements

AU - Kanarachos, Stratis

AU - Blundell, Mike

AU - Haas, Olivier

AU - Kalliris, Michalis

AU - Kotsakis, Rigas

PY - 2019/5/24

Y1 - 2019/5/24

N2 - —Precipitation can adversely influence road safety. Slippery road conditions have traditionally been detected using reactive methods requiring considerable excitation of the tire forces. Alternatives rely on non-contact methods such as vision, sound or ultrasonic sensors. This study proposes a cost-effective wet road conditions detection method based on acoustic measurements for urban and highway driving. It compared the performance of a range of machine learning algorithms to classify the road condition based on the audio features calculated using octave-band frequency analysis. The approach was evaluated experimentally using data collected from a vehicle instrumented with a microphone, GPS and CAN bus data logger. Support Vector Machines using Quadratic and Cubic kernels, as well as Logistic Regression performed better compared to other machine learning-based methods

AB - —Precipitation can adversely influence road safety. Slippery road conditions have traditionally been detected using reactive methods requiring considerable excitation of the tire forces. Alternatives rely on non-contact methods such as vision, sound or ultrasonic sensors. This study proposes a cost-effective wet road conditions detection method based on acoustic measurements for urban and highway driving. It compared the performance of a range of machine learning algorithms to classify the road condition based on the audio features calculated using octave-band frequency analysis. The approach was evaluated experimentally using data collected from a vehicle instrumented with a microphone, GPS and CAN bus data logger. Support Vector Machines using Quadratic and Cubic kernels, as well as Logistic Regression performed better compared to other machine learning-based methods

KW - acoustic measurements

KW - wet road surface detection

UR - https://www.scopus.com/pages/publications/85067091619

U2 - 10.1109/ICMECH.2019.8722834

DO - 10.1109/ICMECH.2019.8722834

M3 - Conference proceeding

VL - (In-press)

SP - 265

EP - 270

BT - Proceedings - 2019 IEEE International Conference on Mechatronics, ICM 2019

PB - IEEE

T2 - IEEE 2019 International Conference on Mechatronics

Y2 - 19 March 2019 through 21 March 2019

ER -

Machine learning algorithms for wet road surface detection using acoustic measurements

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this