Graduation date: 2008Normal acceleration characteristics of passenger cars provide valuable information for various aspects of transportation engineering, including use in the geometric designs of intersections, freeway ramps, passing lanes, acceleration lanes, turning bays, or in the development of traffic simulation or fuel consumption models. However, the data used in previous studies are either outdated or characterized by limitations from data collecting technologies and methodologies. Therefore, the study results may not represent the accelerating behavior of current drivers or modern vehicles in terms of time frame, prediction accuracy and study scopes. This thesis presents an acceleration process study based on a current database collected by invehicle Global Positioning System (GPS) technology over an extended period of time. The acceleration trip data used in this study provides a better representation of real world natural driving behaviors. The author of this thesis developed a third-order polynomial model to describe speed profiles of the leading acceleration process at signalized intersections on arterial roadways. The model also investigates roadway physical feature effects and driver-vehicle effects on the acceleration process. The roadway physical features include horizontal alignments, intersection layouts, and number of through travel lanes. The driver-vehicle effect is modeled as the random effect while the study interest is on the driver population instead of the sample drivers in the data.