Network Screening for Travel Time Unreliability Considering Spatio-Temporal Context

Abstract: Travel time unreliability on roadway networks and their mitigation continue to be increasingly strong focus areas for many transportation agencies. While motorists and transportation agencies are often interested in improving corridor and network travel time reliabilities, the predominant state-of-the-art methods available to practitioners focus on individual hotspots. This research developed the "Top 20-20" method – a novel approach to network screening that evaluates hotspots (relatively shorter links) in conjunction with their spatiotemporal dynamics. The methodology also incorporated techniques for concurrent analysis of causal events at both discrete link levels as well as their overarching corridors. Applied to an expansive 2800-mile limited access facility network in Virginia with four years of hourly aggregated probe speed data, the proposed method underwent qualitative validation and established its robustness. The research demonstrates the value of a holistic corridor perspective beyond individual hotspots. While complementing traditional methods, the proposed Top 20-20 methodology offers new capabilities for mapping traffic patterns, understanding spatiotemporal correlations, and identifying potential causes for unreliable travel times. An expert panel recognized various applications for these methods, including the identification of travel time reliability needs for long-range planning, and the development and implementation of operational strategies for improving travel time reliability on corridors and networks.