Integrity of estimates of the two-fluid model and gender impacts

Abstract

This paper summarizes a research study to develop a methodology for utilizing naturalistic Global Positioning System (GPS) driving data for two-fluid model estimation. The two-fluid vehicular traffic flow model describes traffic flow on a street network as a mix of stopped and running vehicles. The parameters of the model essentially represent ‘free flow’ travel time and the level of interaction among vehicles. These parameters have traditionally been used to evaluate roadway networks and corridors with partially limited access. However, the two-fluid model has been found to be a direct result of driver behavior, and can also be used to represent behavioral aspects of driver populations, e.g., aggressiveness, passiveness, etc. Through these behavioral aspects they can also be related to safety on roadways. Due to which the two-fluid model can be considered to be a safety footprint for a particular road or individual driver. Due to which it is critical to understand factors that influence the two-fluid model. In this study, two-fluid models were estimated using naturalistic driving data collected with GPS data loggers in San Luis Obispo (SLO), California. Linear referencing in ArcMap was used to link the GPS data with roadway characteristic data for each element of the roadway network. The linear referencing methodology is the key to relate the GPS driving data with the elements of roadway network. This study explores two fundamental questions: (1) how sensitive are the estimates of the two fluid parameters to various samples? This question is fundamentally important to help define the integrity of the two-fluid model for planning and operational purposes. To this end we use a random sampling approach to address this question. (2) Are there behavioral differences across gender? This provides important behavioral insights on driving behavior across gender. Significant differences were observed between male and female drivers, with female drivers being more aggressive.