Wyoming Low-Volume Roads Traffic Volume Estimation

This research project developed low-cost and effective traffic volume prediction models for rural low-volume roads in Wyoming. The models were developed to support a wide variety of design, planning, and management functions on both state and county road networks. A review of existing methods of traffic volume estimation in other states was carried out. Two main model types were identified - regression models and travel demand models.;The study developed the two model types for Wyoming and recommended the best model for implementation. Two regression models were developed, a linear regression model and a logistic regression model. Each of the regression models was developed using data from 13 randomly selected counties in Wyoming and nine counties were used to verify the utility of the models. The linear regression model had an R square of 64.3% and was verified to be a good predictor of traffic volumes across Wyoming. The logistic regression model was also verified and shown to have a prediction accuracy ranging from 78% to 89%. The travel demand model (TDM) was developed using standard factors and trip rates in the NCHRP Report 365. The TDM was implemented for four south eastern counties in Wyoming. The model was then verified and calibrated by comparing actual traffic volumes to those generated by the model. The calibrated model had percentage root mean square error (%RMSE) and R square values of 50.33% and 74% respectively. The dissertation compared the three models with respect to cost-effectiveness, ease of use, and accuracy and recommended the travel demand model for implementation across Wyoming. The regression models were recommended for applications requiring quick traffic volume estimations and for which lower levels of accuracy are acceptable.