Road safety and maintenance strategy optimization depend on accurate pavement surface friction prediction. In order to predict the Friction Number for Continuously Reinforced Concrete Pavement (CRCP) sections using data taken from the Long-Term Pavement Performance (LTPP) database, this study presents a hybrid machine learning framework that combines Gradient Boosting Machines (GBMs) with Genetic Algorithm (GA) optimization. Twenty input variables from the structural, climatic, traffic, and performance categories were used in the analysis of 395 data points from 33 CRCP sections. With a mean Root Mean Squared Error (RMSE) of 3.644 and a mean R-squared (R2) value of 0.830, the GA-optimized GBM model outperformed baseline models such as non-optimized GBM, Linear Regression, Random Forest, Support Vector Regression (SVR), and Artificial Neural Networks (ANN). The most significant predictors, according to sensitivity analysis, were AADT, Total Thickness, Freeze Index, and Pavement Age. The marginal effects of these variables on the expected friction levels were illustrated using partial dependence plots (PDPs). The results show that the suggested GA-GBM model offers a strong and comprehensible instrument for forecasting pavement friction, with substantial potential for improving safety evaluations and maintenance scheduling in networks of rigid pavement.
Alnaqbi et al. (Thu,) studied this question.