Date of Award
Doctor of Philosophy
Lee D. Han
Lee D. Han; Asad J. Khattak; Hyun Kim; Russell Zaretzki
Crowdsourcing has been a prevailing fashion of data acquisition thanks to the proliferation of smart devices and wireless communication techniques. In the transportation field, emerging crowdsourcing data could potentially facilitate traffic operation and management. A motiving example is Waze, which is a navigation app that allows Waze users (Wazers) to report different types of traffic incidents e.g., accidents, jams, potholes, disabled vehicles, and so on. Crowdsourced data has many advantages over existing fixed location sensors: cost-effective, extensive geographic coverage, high resolution, and real-time. Moreover, it assembles the human’s perception and judgment of traffic incidents, which is more straightforward compared to raw data output from detectors.
This dissertation centers on exploring the application of crowdsourced Waze data in pavement maintenance, incident management, and truck traffic monitoring. Specifically, the first chapter exploited pothole reports to evaluate the pothole spatiotemporal occurrence and hotspot. The pothole reports were compared to the pothole work request and work record from the pavement management system. The second chapter investigated both the association and causation between potholes and flat tire frequency on highways and discussed the spatiotemporal heterogeneous effect of potholes on flat tire frequency. The third chapter made use of accident and jam reports to estimate traffic recovery time after the crash is cleared. This chapter developed a spatiotemporal statistics model to reveal how factors influence the total crash-induced impact duration. Lastly, the dissertation presented a method to estimate the long truck volume by harnessing crowdsourced speed data as well as traffic detector data. In summary, this dissertation contributes to expanding the potential uses of crowdsourcing for managing pavement and incidents, as well as monitoring traffic.
Gu, Yangsong, "Mining Crowdsourced Data for Pavement Maintenance, Incident Management, and Traffic Monitoring. " PhD diss., University of Tennessee, 2023.
Available for download on Wednesday, May 15, 2024