Lead Presenter Biography

Dr. Shoupeng Tang has more than 16 years of experience in a wide variety of transportation planning and engineering projects. In the past 10 years, Shoupeng has been focusing on travel forecasting and data modeling related research and projects. With his great knowledge in transportation, programming, and data modeling, Shoupeng has worked with National Science Foundation, Federal Highway Administration (FHWA), Texas Department of Transportation (TxDOT), over twenty Texas Metropolitan Planning Organizations (MPO) and TxDOT districts, City of Austin, public transportation agencies and private consulting firms to improve the transportation systems for a better life.

Co-Authors

Presentation Description

In travel demand models (TDM), a volume delay function (VDF) is used in the traffic assignment stage to calculate the link travel times as a function of the volume-to-capacity ratio of each network link. The ongoing research efforts have led to the development of several variations of VDFs with the Bureau of Public Roads (BPR) function being a commonly used function. A handful of VDF parameter estimation methodologies have also been proposed. Numerous research projects have also proposed specific values for different VDF parameters to better replicate the observed field observations by roadway functional class and area type.

This study focuses on improving the traffic assignment results through calibrating the BPR VDFs implemented in the TexPACK platform based on observed and passive datasets. TexPACK is the TDM environment adopted by the Texas Department of Transportation (TxDOT) for small- and medium-sized Metropolitan Planning Organizations (MPO) in Texas and currently used by 19 MPOs (aka, study areas) across the state. With the improvements in computing power and passive big-data sources in the past decade, the scope of this type of analysis can be expanded to a larger study area or the entire state instead of just focusing on a single corridor.

This investigation proposes a networkwide VDF calibration methodology that utilizes statewide observed traffic counts, acquired through TxDOT, and INRIX speed data by small time intervals to calibrate the VDF parameters. The effort led to the successful development of a complete set of procedures to calibrate the VDF parameters in the BPR function. The process also includes developing a model in ArcGIS platform for transferring the count data to the TDM network, and an application for cleaning/smoothing the speed data. These procedures are reasonable, repeatable, and applicable across different study areas. The procedures have been tested on one time-of-day model and six daily models in Texas. All tests have shown positive improvements in terms of model validation results. 

Presentation File