Evaluation of Frictional Pavement Resistance as a Function of Aggregate Physical Properties

Frictional resistance or skid resistance on pavement surfaces can quantify the safety aspects of road users and impact the efficiency of their travel. Skid resistance primarily depends on the textural characteristics of the pavement surface, which is mostly affected by aggregate surface characteristics and aggregate gradation. Many studies were conducted in the past to identify causative factors that affect skid resistance. Few models were also proposed for predicting asphalt pavement skid resistance as a function of various parameters such as mixture characteristics, pavement temperature, traffic level, and traffic speed exclusive of the theoretical water depth and film on the pavement surface to meet the actual conditions of pavement, especially in wet weather conditions. By doing experimentation on two types of modified binder [polymer-modified binder (PMB) and crumb rubber–modified binder (CRMB)], three different types of aggregate sources, and two types of mixes (open-graded friction course and dense-graded bituminous course), frictional properties of the mixes have been determined in terms of British pendulum number (BPN). Water levels were varied in a range of 0–1 mm on the slab specimen to find the consequence of BPN, especially in wet weather conditions. In this paper, symbolic regression with genetic programming was used to develop the empirical model for BPN by using experimental observations. The efficiency of the model has been tested against the coefficient of determination and Nash-Sutcliffe index. The developed model has been found to generalize the underlying relationships among various interdependent causative factors and is able to predict values with high accuracy.