Predicting Pavement Functional Performance Using Machine Learning: A Case Study of the International Roughness Index

Document Type : Research Paper

Authors

1 Department of Civil and Environmental Engineering, Amirkabir University of Technology (Tehran Polytechnic ), Tehran, Iran

2 Department of Civil and Environmental Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran

Abstract

Pavement networks play a vital role in national transportation infrastructure and economic growth by enabling the safe, rapid, and economical movement of goods, services, and people. The quality of these networks is heavily influenced by various types of distress; therefore, accurate prediction of such distress is essential for effective pavement management. Furthermore, machine learning models have recently demonstrated significant potential in modeling pavement performance. This study aims to employ machine learning models to predict the functional condition of pavements, specifically the International Roughness Index (IRI). The research data were extracted from the Long-Term Pavement Performance (LTPP) database managed by the U.S. Federal Highway Administration. The dataset comprises 4,453 records related to pavement structure and construction, weather, traffic, and pavement performance, encompassing 12 effective variables. Seven machine learning algorithms—Decision Tree, Random Forest, XGBoost, Gradient Boosting, K-Nearest Neighbors, Support Vector Regression, and Artificial Neural Network—were used to predict the IRI. Model performance was evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and the Coefficient of Determination (R²). Comparative analysis revealed that the XGBoost and Random Forest algorithms outperformed the others in predicting IRI, with MAE values of 0.17 and 0.18 and R² values of 0.73 and 0.74, respectively. The model developed in this research can serve as a precise and practical tool in pavement management systems for timely roughness prediction and optimizing maintenance programs and budget allocation.

Keywords

Main Subjects

Files

History

  • Receive Date: 19 November 2025
  • Revise Date: 18 February 2026
  • Accept Date: 20 February 2026

Share

How to cite

Statistics