Assessing the Relationship between Speed Behavior and Road Features of a Semi-Urban Mountainous Road using Multivariate Regression

Abstract

The design speed of a highway is used to design its geometric features, which in turn affect vehicle operating speeds. However, operating speeds also depend on driver behavior, which is not uniform and is affected by the highway's features. This research examines this relationship by first applying multivariate linear regression and backward step regression to speed behavior - modeled as the difference between the operating and design speeds - and road feature data from a semi-urban mountainous highway in Nepal, and then develops statistical models for predicting speed behavior for different vehicle categories. The results showed that all geometric features had a significant effect on the speed behavior, and the prediction model for the all-vehicle category using backward step regression was found to be the most accurate. The developed models can be applied to predict the speed behavior for other road sections and support safety management and improvement works.