Automated Vehicle Detection Systems Data as a Promising Approach for Roadway Construction Planning: The Case of Dallas, Texas, Highway System

Abstract

Construction activities on roadway infrastructure have adverse impacts on traffic safety and mobility because of often needed lane and shoulder closures. Existing approaches for planning roadway construction operations primarily rely on general guidelines that do not consider the unique traffic demand patterns of individual roadways. This study proposes utilizing data collected through automated vehicle detection systems (AVDS)— big AVDS data—to enable more informed, data-driven construction planning. To operationalize this approach, the study develops an integrated framework for assessing quality (e.g., validity, completeness) and aggregating AVDS data at various temporal and spatial levels. The framework is demonstrated using big AVDS data for the Dallas, Texas, highway network. The analyses demonstrate how integrating data completeness with performance measures’ behavioral patterns can provide improved understanding of the accuracy of information provided by the AVDS data. The traffic pattern analyses reveal evening peaks that are greater than the morning peaks, remaining so past the typical off-peak time according to the general lane closure guidelines. These findings demonstrate that historically used time periods for construction planning may underestimate demand at traditionally off-peak times. Therefore, the framework demonstration emphasizes the significance of using granular traffic data that considers the specifics of individual roadways to better inform construction planning. The study contributes to advances in construction and engineering management by (i) extending existing roadway construction planning approaches through using higher granular traffic data to better inform roadway construction planning, and (ii) developing a much needed integrated quality assessment and aggregation framework that can operationalize this use.