Calibration of Controlling Input Models for Pavement Management System

The Oklahoma Department of Transportation (ODOT) is currently using the Deighton Total Infrastructure Management System (dTIMS™) software for pavement management. This system is based on several input models which are computational backbones to develop maintenance and rehabilitation plans for pavements. Some of the major input models include the classification of pavement families, deterioration curves, and effectiveness of various treatment options. These major input models are currently in active use without any thorough validation using actual pavement condition assessment data. Validation and calibration of existing input models for pavement management systems (PMS) has been one of the major technical goals by the pavement management unit of ODOT for many years. ODOT now has about 16 years of pavement condition assessment data, which provides a rich time series dataset. This research project will use the proven Knowledge Discovery in Database (KDD) approach to investigate pavement condition assessment data in a structured manner in order to evaluate the performance of current input models and if necessary, develop new models or calibrate the existing models for more accurate and reliable planning for pavement maintenance and rehabilitation activities. The performance of newly developed or calibrated input models will be compared with the performance of current input models. The successful completion of this research project meets the immediate technical need of the pavement management unit. The data driven models developed in this project provides confidence to the pavement management team in developing short-term and long-term pavement management strategies and realistic pavement budget estimation and allocation. One of the primary outputs of this research project is a spreadsheet-based tool that assists pavement management engineers in updating the input models in the current PMS; thus, the output of this project will be immediately available to ODOT. The results of this project will also be able to answer skeptical questions about the returns on continuous pavement data collection investments of ODOT.