Estimating Pavement Temperature Using Machine Learning Models Based on Climatic Data

Abstract: Pavement temperature is a key determinant of pavement performance, influencing material properties, deformation, fatigue, and overall durability. Accurate prediction of pavement temperature enables more effective maintenance strategies, improved pavement design, and enhanced management of infrastructure. In this study, we propose a machine learning-based approach to estimate pavement temperature using secondary climate data, without the need for direct pavement temperature measurements. The objective is to model the relationship between various climatic variables and pavement temperature, leveraging machine learning algorithms for enhanced predictive accuracy.

Climatic data, including air temperature, solar radiation, wind speed, relative humidity, and cloud cover, are obtained from meteorological databases. These factors directly affect pavement temperature and are used as input features for the machine learning models. Algorithms such as artificial neural networks (ANNs), random forests (RF), and gradient boosting machines (GBMs) are employed to model this relationship. The models are trained on historical climate data and validated using a test dataset to ensure their predictive capabilities.

The study involves rigorous data preprocessing, including normalization, feature selection, and cross-validation, to optimize the model’s performance. Evaluation metrics such as mean absolute error (MAE) and root mean square error (RMSE) are used to assess model accuracy. Results indicate that the machine learning models can accurately predict pavement temperature based solely on climate data, outperforming traditional empirical models typically used in pavement engineering.

This research demonstrates that machine learning algorithms offer a reliable, cost-effective solution for estimating pavement temperature, particularly in regions where direct measurements are unavailable. The findings suggest that integrating machine learning into pavement management systems can significantly improve the accuracy of temperature-related predictions, aiding in the optimization of pavement design and maintenance decisions.