Traveler Perception of Transportation System Performance Using Kernel Density Estimation to Prioritize Infrastructure Investments

There is worldwide interest of transportation professionals to quantify traveler perceptions of system performance, ascertaining how such perceptions can differ from objective performance by traditional metrics. Such perceptions include the operational variability of vehicle travel times across hours and days of the week within highway transportation networks. Improvements to transportation infrastructure are informed by performance metrics; however, traditional methods evaluate delays based on deviations from a discrete or ideal condition. In this paper, we measure traveler perception with a novel approach of evaluating delays as deviations from the speed value with the maximum kernel density estimate (KDE). This approach provides a foundation for a risk-based multicriteria framework to inform stakeholders of appropriate reliability and safety mitigation methods. Recent advances in vehicular volume and speed data collection provide the disaggregate traffic data that depicts the variability across disparate time periods. The framework demonstrated in this paper informs enterprise operators and transportation agencies with new perspectives of relative congestion and infrastructure investment planning.