TY - GEN
T1 - Approaching vehicle detection method with acoustic analysis using smartphone for elderly bicycle driver
AU - Kawanaka, Shogo
AU - Kashimoto, Yukitoshi
AU - Firouzian, Aryan
AU - Arakawa, Yutaka
AU - Pulli, Petri
AU - Yasumoto, Keiichi
N1 - Funding Information:
This work was supported by JSPS KAKENHI Grant Number 15H05708.
Publisher Copyright:
© 2017 IPSJ.
PY - 2018/4/2
Y1 - 2018/4/2
N2 - More than 60 percentage of fatal accidents while riding a bicycle is caused by elderly people over 65 years old. The main cause is the detection delay of approaching vehicle caused by the decrease of cognitive function due to aging. In this paper, we propose an approaching vehicle detection method using a smartphone aiming to support bicycle operation to prevent elderly people from fatal accidents while riding a bicycle vehicle. Among various sensors embedded in a smartphone, we focus on microphone as the most suitable sensor for detecting an approaching vehicle. We collected sound data in a real environment and created an approaching vehicle detection model by using machine learning. Finally, as a result of accuracy evaluation with 10-fold cross-validation, we confirmed that it can detect approaching vehicle with an average F-value of 97.4 [%].
AB - More than 60 percentage of fatal accidents while riding a bicycle is caused by elderly people over 65 years old. The main cause is the detection delay of approaching vehicle caused by the decrease of cognitive function due to aging. In this paper, we propose an approaching vehicle detection method using a smartphone aiming to support bicycle operation to prevent elderly people from fatal accidents while riding a bicycle vehicle. Among various sensors embedded in a smartphone, we focus on microphone as the most suitable sensor for detecting an approaching vehicle. We collected sound data in a real environment and created an approaching vehicle detection model by using machine learning. Finally, as a result of accuracy evaluation with 10-fold cross-validation, we confirmed that it can detect approaching vehicle with an average F-value of 97.4 [%].
UR - http://www.scopus.com/inward/record.url?scp=85049628485&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85049628485&partnerID=8YFLogxK
U2 - 10.23919/ICMU.2017.8330069
DO - 10.23919/ICMU.2017.8330069
M3 - Conference contribution
AN - SCOPUS:85049628485
T3 - 2017 10th International Conference on Mobile Computing and Ubiquitous Network, ICMU 2017
SP - 1
EP - 6
BT - 2017 10th International Conference on Mobile Computing and Ubiquitous Network, ICMU 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 10th International Conference on Mobile Computing and Ubiquitous Network, ICMU 2017
Y2 - 3 October 2017 through 5 October 2017
ER -