Abstract:It has been reported that the pavement performance predicted by the current mechanistic-empirical pavement design shows low or no sensitivity to subgrade and unbound layers. This issue has raised wide attention. Targeting this problem, this paper summarizes the process used by the authors to find better models of the influence of subgrade and unbound base course layers on the performance of flexible and rigid pavements. A comprehensive literature review is first conducted and the findings are categorized. It is found that the resilient modulus, permanent deformation, shear strength, and erosion are key factors. In particular, the properties that provide greater sensitivity are (1) the moisture-dependency of the modulus, shear strength, and permanent deformation; (2) stress-dependency of the modulus and permanent deformation; and (3) cross-anisotropy of the modulus. A number of unbound layer/subgrade models have been located and categorized. Three criteria are developed to identify the candidate models in terms of the degree of susceptibility, degree of accuracy, and ease of development. The first two criteria are used to evaluate the collected unbound layer/subgrade models, while associated development and implementation issues are planned as subsequent work. Two models that the authors previously developed are selected as examples to illustrate the improvement of the performance prediction, including the moisture-sensitive, stress-dependent, and cross-anisotropic modulus model for unbound layers and stress-dependent mechanistic-empirical permanent deformation model for unbound base layers. These two models are verified through laboratory tests and numerical simulations. Moreover, they are compared to their counterparts in the AASHTOWare Pavement ME Design. The advantages of accuracy and sensitivity to the operational conditions (e.g. moisture, traffic stress, and load-induced/particle-induced anisotropy) are obvious. In addition to these two models, the development of the shear strength model and erosion model are sketched. The candidate models need further development and implementation, which address issues such as hierarchical inputs, calibration/validation, and implementation. These are the on-going and planned work on this topic to better incorporate the influence of subgrade and unbound layers so as to contribute to the improvement of pavement designs.

An interpretation framework of machine learning models: prediction of pavement long-term performance

Mechanistic-empirical Models for Better Consideration of Subgrade and Unbound Layers Influence on Pavement Performance

Determination of Flexible Pavement Deterioration Conditions Using Long‐Term Pavement Performance Database and Artificial Intelligence‐based Finite Element Model Updating

A New Long-Term Aging Model for Asphalt Pavements Using Morphology-Kinetics Based Approach

Pavement Roughness Prediction Using Explainable and Supervised Machine Learning Technique for Long-Term Performance

Creating Rutting Prediction Models through Machine Learning Techniques Utilizing the Long-Term Pavement Performance Database

Rutting prediction and analysis of influence factors based on multivariate transfer entropy and graph neural networks

A hybrid framework for asphalt pavement rutting prediction modeling and influencing factors analysis based on multilevel wavelet decomposition and transfer entropy

An ensemble learning model for asphalt pavement performance prediction based on gradient boosting decision tree

Prediction of International Roughness Index Based on Stacking Fusion Model

Imporved Model for Pavement Performance Prediction Based on Recurrent Neural Network Using LTPP Database

A novel approach for asphalt pavement invisible distress classification prediction by machine learning algorithms

Machine Learning for Prediction of the International Roughness Index on Flexible Pavements: A Review, Challenges, and Future Directions

A Data-Driven Rutting Depth Short-Time Prediction Model with Metaheuristic Optimization for Asphalt Pavements Based on RIOHTrack

Investigating long-term performance of flexible pavement using Bayesian multilevel models

Enhancing winter road maintenance with explainable AI: SHAP analysis for interpreting machine learning models in road friction estimation

Performance Prediction of Expressway Pavement in High Maintenance Level Areas Based on Cosine Deterioration Equation: A Case Study of Zhejiang Province in China

Performance of Machine Learning Algorithms in Predicting the Pavement International Roughness Index

Machine Learning Approach to Predict International Roughness Index Using Long-Term Pavement Performance Data

Prediction of Decay of Pavement Quality or Performance Index Based on Light Gradient Boost Machine

A Surface Texture Prediction Model Based on RIOHTrack Asphalt Pavement Testing Data