Auto-adaptive pavement performance prediction methodology.

This paper was presented at the `Strategies for enhancing pavement performance' session. This paper describes an auto-adaptive modeling approach to predicting pavement performance that provides site-specific adaptation as well as generic feedback for updating default model coefficients. The predictive methodology is applicable to a wide variety of contexts including prediction of roughness based indices, composite rating indices and/or individual pavement distresses. It can be used for both the network and project level for analysing maintenance and rehabilitation alternatives. The modeling approach is based on the use of historical performance data to develop model coefficients adapted to individual analysis sections, as well as for the feedback analysis used to update the table of default model coefficients. (A)