Degree Type

Creative Component

Semester of Graduation

Spring 2020

Department

Civil, Construction, and Environmental Engineering

First Major Professor

Dr. Anuj Sharma

Degree(s)

Master of Science (MS)

Major(s)

Civil Engineering

Abstract

Traffic incidents impact traffic system components and have a major contribution to congestion, delays, and secondary incidents. In order to avoid these issues, all traffic control systems are trying to detect incidents in a timely manner using different sources of traffic monitoring. This research has developed a comparison method among different sources of real-time incident detection in terms of temporal and spatial characteristics. Incident reports from Waze and ATMS and generated automated incidents from the Inrix, and Wavetronix were used for this analysis. The result shows that all data sets have a similar pattern in recording incidents. The number of incidents recorded by Waze, Wavetronix, and Inrix tracks each other closely during different times of the day. ATMS is a validated set of incidents reports, but still, there are more incidents that were not recorded in ATMS. A state-of-the-art matching function is built to get the number of overlapping incidents from all four data sets. Waze and Inrix have the highest overlapping with ATMS, about 57%, which affirms their reliability. The matching function also was used to get the estimated contribution of each source to ATMS. Due to the lack of validation source, there will be some false alarms in the total number of contributions that were not identified.

Copyright Owner

Khalilzadeh Fatemeh

File Format

PDF

Share

COinS