Skip to main content
Open Access Publications from the University of California

Modeling Pavement Performance by Combining Field and Experimental Data


The accurate prediction of pavement performance is important for efficient management of the surface transportation infrastructure. By reducing the error of the pavement deterioration prediction, agencies can obtain significant budget savings through timely intervention and accurate planning.

The goal of this research was to develop a methodology for developing accurate pavement deterioration models to be used primarily for the management of the road infrastructure. The loss of the riding quality of the pavement was selected as the performance indicator. Two measures of riding quality were used: serviceability (Present Serviceability Index, PSI) and roughness (International Roughness Index, IRI).

An acceptable riding quality is important for both the road user and the goods being transported. Riding quality affects the comfort of the user for whom the road is provided, and the smoothness with which goods are moved from one point to another. The vehicle operating costs and the costs of transporting goods increase as the road riding quality deteriorates. These costs are often one order of magnitude more important than the cost of maintaining the road to an acceptable level of service.

The initial incremental models developed in this dissertation predict serviceability as a function of material properties, pavement structural characteristics, traffic axle configuration, axle load, and environmental variables. These models were developed applying nonlinear estimation techniques using an experimental unbalanced panel data set (AASHO Road Test). The unobserved heterogeneity among the pavement sections was accounted for by using the random effects approach.

The serviceability models were updated using joint estimation with a field panel data set (MnRoad Project). The updated model estimates riding quality in terms of roughness. This was possible by applying a measurement error model to combine both data sources.

The main contribution of this research is not the development of a deterioration model itself, but rather the demonstration of the feasibility of using joint estimation and its many advantages, such as: (i) identification and quantification of new variables, (ii) efficient parameter estimates, (iii) bias identification and correction, and (iv) use of a measurement error model to combine apparently incompatible data sources.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View