Dynamic Panel Data Model for Predicting Performance of Asphalt Concrete Overlay

Abstract

In order to identify the improvement needs on transportation facilities and the associated budget, the transportation agencies need to have a reasonable estimate of the facilities’ future condition. Using asphalt concrete overlay over fractured portland cement concrete pavement as an example, this paper shows how to use the existing pavement condition and other recorded information in a pavement management system to predict the future condition of pavements. A comprehensive pavement performance indicator, condition points, was used as the response variable of the prediction model. A dynamic panel data prediction method was employed to make the prediction model reflect the influence of the existing pavement condition, exogenous variables, and the heterogeneity of different pavement sections. The predicted condition points were compared with the actual values and the results indicated that the prediction accuracy was reasonable.