Comparative Analysis of Entry Capacity Models for Multilane Roundabouts under Mixed Traffic Flow Conditions

Abstract

Evaluation of roundabout capacity holds significant importance because it directly affects the performance, safety, and environmental issues at the intersections. Several researchers have presented models for estimation of the entry capacity of roundabouts, but most of them have been limited to traffic conditions in developed countries having proper lane discipline, and upto single-lane and two-lane roundabouts only. There is dearth of such studies for developing countries, such as India, where traffic conditions are entirely different from developed countries. This study aims to evaluate the capacity of both single-lane and multilane roundabouts in handling different levels of heterogeneous traffic flow. Data were collected at 10 roundabouts spread across three Indian cities. All the roundabouts have different geometrical features and lane configurations, thereby ensuring the applicability of this study across different roundabouts. The formation of queues was used to indicate that the approach is operating at its maximum capacity. Four different models (linear regression model, non-linear regression model, VISSIM simulated model, and calibrated Highway Capacity Manual [HCM] model) were developed for each roundabout, and results were compared with the field capacity values. Careful consideration was given to statistical characteristics of the data and different statistical tests were conducted to evaluate the reliability of these models. A separate set of roundabouts was used to test and validate the models. Results reveal that a non-linear regression model, based on both the geometrical and traffic flow characteristics, outperforms all other models. Additionally, sensitivity analysis was performed to check the effect of various parameters on the roundabout capacity.