Enhanced Operating Speed Prediction Models for Horizontal Curves on Non-Urban Two-Lane Roads: A Case Study from South India

The inconsistency in geometric parameters of road alignment significantly influences individual vehicle speeds, contributing to increased crash occurrences. The lack of uniformity and accuracy in existing operating speed models has led to improper selection of geometric parameters, impacting road design consistency and vehicle operating speed. Furthermore, current models are limited to passenger cars and often consider only a narrow set of geometric variables.<br /> <br /> This study addresses the limitations of poor accuracy and uniformity in existing models by including all types of vehicles and incorporating a comprehensive range of geometric variables influencing speed. Traffic data from 261 curves and 261 tangents are used to develop refined speed prediction models. A mixed-effect modelling approach is employed to account for the clustered complexity of the model variables. The proposed mixed-effect models account for the heterogeneous nature of traffic conditions in India. These models demonstrate the ability to compute operating speeds with high accuracy, especially when compared to existing models developed under homogeneous traffic conditions with limited data. By considering a broader range of vehicles and geometric variables, this work aims to contribute to more reliable and applicable operating speed models for diverse traffic scenarios.