Semester

Spring

Date of Graduation

2006

Document Type

Thesis

Degree Type

MS

College

Statler College of Engineering and Mineral Resources

Department

Civil and Environmental Engineering

Committee Chair

David R. Martinelli.

Abstract

This study documents the development of a methodology and models to forecast truck traffic volumes on Interstate Highways at a selection of Permanent Automatic Traffic Recorder (PATR) sites. The models were developed using data collected over a period nine years (1995 to 2003) from sixteen permanent count stations located throughout the state. Eight sites were ultimately utilized, five along rural interstate highways and the others from urban interstate highways. Model development was based on the time series method, using two techniques: regression analysis and the growth factor technique. Both were analyzed and compared in order to select the most reliable technique to be used in the forecasting procedure. To further understand changes in truck traffic patterns, traffic was grouped according to the FHWA vehicle classification scheme. Models were developed for each site and for every truck classification in these sites as well. Due to the smaller effect of demographic characteristics on interstate highways models; these models were performed using as a predicted variable: the Annual Average Daily Truck Traffic data obtained directly from the counters, and time period as the unique independent variable. Validation was conducted using the coefficient of variation to measure the statistical significance of the results obtained. Further validation of models was conducted by the coefficient of regression, and by comparison between the based trends data with the predicted models. In the course of the study, regression models resulted as the appropriate predictor technique to be used at interstate highways. Models, growth factors and figures are reported by every site and truck classification, detailed tables containing these factors are presented in the report.

Share

COinS