Abstract
State highway agencies are required to establish performance targets for pavement facilities based on national performance measures. Because of the inconsistency and uncertainty in performance measures between the national and state systems, states are facing challenges in utilizing national performance measures to set state performance targets. Furthermore, because of the lack of historical data, it is difficult to establish reliable models to predict future performance measures. In this study, the cross-correlations among the national performance metrics were investigated. The national performance measures were correlated with the state performance indices by a probabilistic method to consider the influence of uncertainty on setting the performance targets. Random forest method was employed to correlate the national performance measures to state performance indices. The results indicated that the international roughness index generally increased with the increase of standard deviation of rut depth. A single performance index was not correlated well with national measures of “poor” condition, whereas the pavement quality index (PQI) correlated well with “good” condition. The accuracy of the classification model with two indices (pavement smoothness index, pavement distress index) was higher than that with a single index (PQI). The probabilistic curves were developed to correlate state performance index and national measures, which can be used for performance target setting. This paper demonstrates that the national measures could be introduced into the state decision-making process by establishing probabilistic relationships between the national performance measures and the state pavement condition indices.
Get full access to this article
View all access options for this article.