This study aims to empirically analyze the structural causes of recurrent traffic congestion in urban street segments and to identify the relationships between road components that influence congestion. While previous research on traffic congestion ha...
This study aims to empirically analyze the structural causes of recurrent traffic congestion in urban street segments and to identify the relationships between road components that influence congestion. While previous research on traffic congestion has primarily focused on traffic volume or demand management, this study utilizes DSRC-based (Dedicated Short-Range Communications) real-world speed data to quantitatively examine the physical and operational characteristics of road segments where congestion repeatedly occurs. Going beyond the simplistic notion of “road expansion = congestion relief,” this research proposes a congestion management strategy based on geometric optimization through detailed analysis of road geometry and intersection operations.
The study area includes 74 arterial road segments in the urban center of Daegu, South Korea. Approximately 700,000 DSRC speed records were used to calculate congestion days, and K-means clustering was applied to classify congestion levels into four categories (non-congested to heavily congested). Using geometric and operational data from each segment, a multinomial logistic regression model was developed with 10 independent variables. Additionally, interaction effects between variables were analyzed using a Z-score standardized interaction model.
The analysis found that factors such as traffic volume per lane, number of signalized intersections, number of turning sections, and lane configuration changes significantly influence congestion levels. Interaction terms—such as traffic × signal density, signal × turning, and signal × bus-only lane—exhibited stronger explanatory power than individual variables. These findings suggest that targeted design and policy interventions, such as identifying geometry-sensitive congestion zones, refining signal timing strategies, and improving lane guidance infrastructure, are essential for effective urban congestion management.