Manuscript received July 8, 2023; revised August 28, 2023; accepted October 2, 2023.
Abstract—Healthy road networks are essential to facilitate economic and social development. Sustaining the integrity of road pavement necessitates having reliable performance models. Such performance models can facilitate evaluating the effect of different physical, environmental, and operational factors on road pavement performance. Hence, this research adopts multiple Machine Learning (ML) algorithms to model the impact of these factors on a composite Pavement Condition Rating (PCR). The PCR is developed using three indicators, namely cracking, rutting, and the International Roughness Index (IRI). This study investigates the implementation of some widely acknowledged ML algorithms, including Artificial Neural Networks, Support Vector Machines, and Bagged Regression Trees to model road pavement performance. Thus, the models are developed and tested using a data set of 302 road sections managed by the Nebraska Department of Roads (NDOR). Also, the deterioration factors are ranked based on their influence on the PCR. Based on the developed models, annual daily traffic (ADT), base layer thickness, and age affect the pavement condition most.
Copyright © 2012-2023 International Journal of Structural and Civil Engineering Research, All Rights Reserved