SPR 14-469: Pavement Performance Modeling

The Transportation Cabinet has been collecting pavement performance information for a number of years, including rideabiliy, condition points, individual distress information and rutting. To fully utilize this information modeling must be performed to utilize this data to predict what future performance levels may be. This information is invaluable when working to establish future highway network rehabilitation needs and treatments. This study would develop performance prediction models for various classes of roadways and pavement types that could then be used on a network level and project level. The primary benefit will be a mechanism to predict future conditions of roadway segments and networks. This will allow better decisions to be made with respect to funding allocation and optimization of the funding which may become available. The objectives of this project are 1. Review current Pavement Management System to determine pre-defined models that may be available. 2. Review available distress and condition data for various types of pavements, interstates, parkways, etc. 3. Evaluate various modeling techniques for performance modeling, Neural Networks, Multiple Regression, Multiple Adaptive Regression Splines, and others. 4. Develop models which can be incorporated into the existing Pavement Management System.