International Roughness Index Prediction Using Various Machine Learning Techniques on Flexible Pavements

Pavement performance is a significant factor for transportation network to devise its strategies and policies regarding construction and maintenance of damaged roads. This can be evaluated using Pavement roughness index that indicates roughness of the pavement. Unfortunately, there is no standard Roughness Index metric in India. This study explores the various models to evaluate the structural parameters of the road. For the same, dataset is taken from Central Road Research Institute of India (CRRII) database. Various machine learning models are applied in the dataset to determine the pavement condition. During the experimental evaluation, it is noticed that Artificial Neural Network and XGB-Regressor yield the best performance by giving the smallest mean absolute chance error.