Travel Time Reliability Models

Two previous Florida Department of transportation (FDOT) research projects )FDOT Contracts BD-545-48 and BD-545-70) on travel time reliability developed models for predicting travel time reliability for freeways, using data from Philadelphia, Pennsylvania. The objectives of this research are to implement the results of these projects to estimate travel time reliability on the Florida Strategic Intermodal System (SIS), to enhance the previously developed models based on Florida data, and to develop a framework for estimating travel time on arterial streets. The first part of the report describes a methodology and respective tools for estimating travel time reliability for a freeway section as well as for a freeway network, and it applies those tools to estimate various travel time reliability measures for a freeway section in South Florida. The second part of the report summarizes the data analysis conducted using data from Jacksonville, Florida to evaluate the travel time estimation models previously developed. It was concluded that the difference between model estimates and field observations are relatively low for non-congested conditions (less than 5%), while they are relatively larger for congested conditions (as high as 23.7%). The discrepancies may be due to model limitations as well as sensor calibration issues. The third part of the report developed a framework for estimating arterial travel time and obtaining the related travel time reliability performance measures. It is recommended that a total of 16 scenarios be used to obtain travel time estimates for arterials. Future work should focus on obtaining field data to calibrate the conceptual models provided in this report. Generally, it is recommended that the methodology developed and illustrated in this report be implemented to estimate travel time reliability for the SIS and evaluate the model predictions for existing conditions. Such an evaluation will identify any atypical conditions for which adjustments to the existing models need to be made and will identify any inconsistencies between predicted and observed travel time. Subsequently, the application can be used to test various project implementations and estimate travel time reliability in the entire SOS, or is specific SIS sections as a function of various operation and design changes.