TY - GEN
T1 - Initial Design of Two-Stage Acoustic Vehicle Detection System for High Traffic Roads
AU - Uchino, Masato
AU - Dawton, Billy
AU - Hori, Yuki
AU - Ishida, Shigemi
AU - Tagashira, Shigeaki
AU - Arakawa, Yutaka
AU - Fukuda, Akira
N1 - Funding Information:
ACKNOWLEDGMENT This work was supported in part by JSPS KAKENHI Grant Numbers JP15H05708, JP17H01741, and JP18K18041 as well as the Cooperative Research Project of RIEC, Tohoku University.
Publisher Copyright:
© 2020 IEEE.
PY - 2020/3
Y1 - 2020/3
N2 - As the adoption of Intelligent Transport Systems (ITS) grows worldwide, so does the need for lost-cost, fast-deployment vehicle detection systems. SAVeD is a low-cost acoustic detection system developed by the authors which works by fitting a curve indicating vehicle passage to a sound map depicting the difference in arrival time of a passing vehicle's sound at two microphones installed on the roadside. This paper expands on the SAVeD method by proposing a Two-Stage Acoustic Vehicle Detection System for use in high-traffic environments, where multiple simultaneously and successively passing vehicles cause interference in the detection process. To solve this problem, the sound map fitting process is divided into two stages: the detection range is narrowed based on information estimated during the Pre-Fitting stage, and neighborhood point extraction is performed during the Post-Fitting stage to improve vehicle detection accuracy. Initial evaluation performed on a four-lane, two-direction road showed a vehicle detection F-measure of 0.63, a 12-point increase over SAVeD.
AB - As the adoption of Intelligent Transport Systems (ITS) grows worldwide, so does the need for lost-cost, fast-deployment vehicle detection systems. SAVeD is a low-cost acoustic detection system developed by the authors which works by fitting a curve indicating vehicle passage to a sound map depicting the difference in arrival time of a passing vehicle's sound at two microphones installed on the roadside. This paper expands on the SAVeD method by proposing a Two-Stage Acoustic Vehicle Detection System for use in high-traffic environments, where multiple simultaneously and successively passing vehicles cause interference in the detection process. To solve this problem, the sound map fitting process is divided into two stages: the detection range is narrowed based on information estimated during the Pre-Fitting stage, and neighborhood point extraction is performed during the Post-Fitting stage to improve vehicle detection accuracy. Initial evaluation performed on a four-lane, two-direction road showed a vehicle detection F-measure of 0.63, a 12-point increase over SAVeD.
UR - http://www.scopus.com/inward/record.url?scp=85092003703&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85092003703&partnerID=8YFLogxK
U2 - 10.1109/PerComWorkshops48775.2020.9156248
DO - 10.1109/PerComWorkshops48775.2020.9156248
M3 - Conference contribution
AN - SCOPUS:85092003703
T3 - 2020 IEEE International Conference on Pervasive Computing and Communications Workshops, PerCom Workshops 2020
BT - 2020 IEEE International Conference on Pervasive Computing and Communications Workshops, PerCom Workshops 2020
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2020 IEEE International Conference on Pervasive Computing and Communications Workshops, PerCom Workshops 2020
Y2 - 23 March 2020 through 27 March 2020
ER -