Methodology to Counter Heterogeneity at Urban Multilane Roundabouts

Abstract

Despite the adoption of roundabouts in developing countries for traffic management and smooth flow, the concept of traffic heterogeneity management at these intersections is relatively unexplored. Estimating passenger car units (PCUs) for different vehicle types to homogenize traffic is a recognized approach worldwide. In this study, vehicles were classified into five categories to provide a meaningful context for the analysis. The study proposes a model to estimate PCU values for different vehicle types, specifically at roundabouts, after identifying the potential influencing factors. The study was based on data collected from 10 roundabouts in three Indian cities, having different physical and geometrical characteristics. The proposed method yielded PCU values that were partially comparable with existing standards. Comparisons with the Indian Highway Capacity Manual (Indo-HCM) revealed that this publication underestimated PCU values for two-wheelers and three-wheelers by 20% and 11% respectively, but overestimated PCU values for big cars and heavy vehicles by 9% and 12%, respectively. Method appropriateness was ascertained from the lack of prior assumptions made about traffic and geometrical characteristics. To prevent unnecessary re-estimation of PCU values across locations, the idea of a heterogeneity counter component (HCC) was proposed. The model was trained using cross validation on a dataset of seven roundabouts, and its performance was evaluated on a separate dataset using mean absolute percentage error. The relationship between PCUs and the geometrical features of roundabouts, and the effect of traffic characteristics on HCC were studied. The findings might be considered while revising relevant codes and manuals (e.g., IRC-65:2017, Indo-HCM).