About OpenKnowledge@NAU | For NAU Authors

Automated Identification of Traffic Detector Malfunctions

Riffle, Katherine Rose (2021) Automated Identification of Traffic Detector Malfunctions. Masters thesis, Northern Arizona University.

[thumbnail of Riffle_2021_automated_identification_traffic_detector_malfunctions.pdf] Text
Riffle_2021_automated_identification_traffic_detector_malfunctions.pdf - Published Version

Download (1MB)


This study develops a novel method of detector performance verification by using collected and transcribed drone video data and comparing it to detector event log data. Previous studies have used existing video cameras to provide a higher quality assessment of detector performance through manual verification and transcription. The same level of assessment was required for the detection areas in this work, as they were to be used to develop an algorithm to monitor detector health over time, presuming they were healthy. However, with no cameras onsite, drones were used as part of an analysis process that included drone video collection, data transcription, and comparison to event log detector outputs through two statistical methods. Drone video recordings were higher quality and provided preferable overhead viewing angles than post-mounted intersection cameras otherwise would have. This scope of this study then turns to developing new methods for evaluating detector health using event-based outputs and existing traffic flow theory. Event-based detector data outputs were used to develop empirical Volume vs. Density curves, per Greenshield’s Fundamental Model. Using integration, these empirical lines were compared with a conceptual Volume vs. Density curve for each detector, generated using average headway data and the posted speed limit. Additionally, detector performance and site information were used to model a predicted Volume versus Density relationship for each detector based upon collected data, which was then compared with the Conceptual line in the same manner as the empirical lines. The outcomes of both of these comparisons were then used to create a database to be used for assessing detector health within the structure of an algorithm. The algorithm is then presented and discussed, followed by directions for future research, lessons learned, and limitations of this work.

Item Type: Thesis (Masters)
Publisher’s Statement: © Copyright is held by the author. Digital access to this material is made possible by the Cline Library, Northern Arizona University. Further transmission, reproduction or presentation of protected items is prohibited except with permission of the author.
Keywords: Traffic detection; Traffic measurement; Drones; Machine reliability; Cameras
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
NAU Depositing Author Academic Status: Student
Department/Unit: Graduate College > Theses and Dissertations
College of Engineering, Informatics, and Applied Sciences > Civil Engineering, Construction Management and Environmental Engineering
Date Deposited: 04 Feb 2022 21:10
Last Modified: 28 Dec 2022 08:30
URI: https://openknowledge.nau.edu/id/eprint/5656

Actions (login required)

IR Staff Record View IR Staff Record View


Downloads per month over past year