Speed-Based Autosegmentation Method for Reporting Moving Ahead for Progress in the 21st Century Act Performance Measures

Abstract

An early step in the process of performing any mobility analysis is the segmentation of the roadway network. Traditional manual segmentation includes reviewing maps and geometric roadway characteristics to segment roadways into logical, similarly behaving segments. This task is time consuming and does not inherently use the actual speed data in the segmentation process. There is a need for an automated procedure to provide a first cut of roadway segments for analysis. The roadway segmentation procedure presented used a comparison of average annual 15-min speeds by day of week to judge whether adjacent roadway links exhibited similar traffic patterns and should be grouped together for mobility analysis. The procedure used relatively simple calculations to provide a single-number criterion indicating the relative degree of similarity between pairs of adjacent or near-adjacent roadway links. Researchers developed an automated data processing framework for autosegmenting freeways by using INRIX speed data and used this processing framework to evaluate the autosegmentation method in comparison with known congested locations in Houston, Texas. Comparison with known congested segments from the Texas Department of Transportation's 100 most congested roadways list showed reasonably good agreement with congested locations in Houston. The methods explained in this paper are particularly useful for transportation agencies interested in segmenting their roadway networks to produce performance measure requirements expected from the Moving Ahead for Progress in the 21st Century Act.