|
Title:
|
HIGHWAY INCIDENT DETECTION BASED ON PROBE CAR DATA |
|
Author(s):
|
Hiroto Akatsuka, Atsuhiro Takasu, Kenro Aihara, Jun Adachi |
|
ISBN:
|
978-972-8939-83-0 |
|
Editors:
|
Miguel Baptista Nunes, Pedro Isaías and Philip Powell |
|
Year:
|
2013 |
|
Edition:
|
Single |
|
Keywords:
|
Automatic Incident Detection, Probe Car Data, Big Data. |
|
Type:
|
Full Paper |
|
First Page:
|
103 |
|
Last Page:
|
110 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
In this paper, we propose a method for automatic incident detection (AID) based on probe car data (PCD). Traffic incidents cause serious traffic congestion, which eventually leads to economic loss and environmental pollution. The early detection of a traffic incident is a very important component of traffic management. Many traffic incidents also occur during disasters, so the early detection of incidents is very important for disaster mitigation and damage limitation. From this perspective, PCD may be regarded as a cost-effective data source. A feature that can be obtained from PCD is different from that of road installation-type sensors. Thus, we investigated a feature that can be obtained from PCD to exploit PCD to the utmost. This feature is an index that represents the speed deviation, which is extracted from each car's temporal and spatial mean speed and used for AID. We applied our method to data collected on days when incidents occurred. The evaluation results showed that our proposed method detected each incident correctly with an average false alarm rate of 0.08% or less. This shows that the method could be practiced effectively. |
|
|
|
|
|
|
