Citation

Yoshida M, Kawasaki Y, Umeda S, Kuwahara M. Transp. Res. Proc. 2018; 34: 179-186.

Copyright

(Copyright © 2018, Elsevier Publications)

DOI

10.1016/j.trpro.2018.11.030

PMID

unavailable

Abstract

Herein, we propose methods that use three-dimensional trajectory data (one-dimensional time × two-dimensional space) acquired from probe vehicles to detect damaged spots on roads caused by natural disasters. Detecting road damages immediately after a disaster is essential to ensure quick and safe evacuations, rescue and relief operations, and efficient road repairs. However, road damages are currently monitored only by closed-circuit television, administrators patrolling the roads, and reports from users of the roads. Consequently, the number of locations that are being monitored is limited, and damage detection is delayed. Our proposed methods automatically estimate the locations of widespread damages on general roads. After analyzing and extracting the features of probe data obtained during a disaster, our proposed methods were verified using the data of past disasters. Our methods can identify anomalous vehicle behaviors and indicate various possibilities for detecting road damages.

Language: en

Keywords

large-scale disaster; probe trajectory data; traffic-fault detection