Validation of the Highway Performance Monitoring System for forecasting levels of traffic

Abstract

This thesis documents the results and studies conducted when determining the accuracy of the Texas Department of Transportation (TXDOT) methodology of forecasting future levels of traffic demand. The data that was used was the Federal Highway Administration's Highway Performance Monitoring System (BPMS), which is a data base containing various information for all public road facilities within the nation. Two sets of data from different test years were obtained from the Texas Transportation Institute (TTI) located on the Texas A&M campus. These data sets were used in the determination of the accuracy of the forecasting method as used by the TXDOT in the submission of the liPMS data. Data from the two test years were obtained from TTI. The test years chosen were 1987 and 1992. The data were estimated using straight line regression as currently used by TXDOT. The data for the year 1987 were projected by the author into the year 1992. Then the actual demand levels recorded in 1992 were compared to these forecasted data points. The studies confirmed that this method of forecasting is not accurate and a new method should be developed. However, overall the project demand levels were much better than anticipated. In each of the classifications there was a pattern to all of the data points: to lie on a straight linejust above the actual data points. The forecasting method generally results in a demand estimate that is too high. If the projected demand values could be lowered by a standard percentage, then the projected data points would be much closer to the actual data points.