Performance Evaluation of Roundabouts under Mixed Traffic Conditions: A Data-Driven Approach

This study presents a comprehensive analysis of roundabout performance under mixed traffic conditions, focusing on various parameters influencing performance effectiveness. Field data from ten roundabouts in three Indian cities, including those with multiple approach lanes, were collected to assess geometric features and traffic characteristics. The study addressed factors such as heterogeneity, gap acceptance behavior, capacity estimation, and level of service (LOS) by adopting a structured, stepwise methodology. Three entry capacity models were developed using different approaches: empirical approach, approach based on gap acceptance behavior, and traffic simulation. A multiple regression-based nonlinear model based on the empirical approach emerged as the most suitable for estimating entry capacity under mixed traffic flow, offering a more robust alternative specifically to existing Highway Capacity Manual (HCM) and simulation models. The HCM model based on drivers’ gap acceptance behavior underestimated the field capacity, whereas the Traffic in Cities Simulation Model (VISSIM) tended to overestimate it. Furthermore, the study introduces two sets of criteria for classifying LOS, based on average vehicle delay and degree of saturation levels. Compared with existing models, these criteria offer improved evaluations by addressing limitations observed in previous methodologies. Utilizing K-means clustering, LOS classification criteria were established that enhance the assessment of roundabout performance. Results demonstrated slight increases in the range of each LOS compared to existing models, with adjustments made to better reflect user experience and roundabout performance. By incorporating the study’s recommendations, urban planners and engineers can optimize design parameters for new roundabouts and evaluate the performance of existing ones to enhance traffic management efficiency.