Report

R101: Automated Roadworks Detection using Next Generation Traffic Data

Authors
Will Hore-Lacy
Anthony Germanchev
Edward Dann
Young Li
June 1, 2020

This project investigated the feasibility of using real-time traffic data from HERE Technologies to detect and validate roadworks activity across Queensland’s road network, with the aim of improving traveller information and supporting auditing of planned roadworks. The research demonstrated that integrating and matching data from the QLDTraffic service and HERE Technologies Traffic API is possible, but highly complex due to the need for geospatial processing and map matching across multiple road geometry datasets. Significant limitations were identified in the coverage, quality, and accuracy of both roadworks and traffic data, restricting the reliability of the approach. Despite these challenges, the project successfully developed an anomaly detection method that produced promising results in some cases and generated valuable insights for future research. Key outcomes included the development of a powerful and adaptable data aggregation methodology, the demonstration that k-means clustering may be applicable to other event types, and an improved understanding of the relationship between traffic impacts and roadwork events. While further improvements in data quality are needed, the project established a strong foundation for future applications of real-time traffic analytics, with increasing probe vehicle coverage expected to enhance accuracy and expand applicability across the road network.

Download ReportDownload Report