טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentLirit Binyamin-Rozenfeld
SubjectRoute Choics Modeling Using Combined Data from
GPS and GIS
DepartmentDepartment of Civil and Environmental Engineering
Supervisor Full Professor Bekhor Shlomo
Full Thesis text - in Hebrew Full thesis text - Hebrew Version


Abstract

An effective route choice forecast is essential for efficient design of future transportation networks, and for improvement and optimization of the existing transportation network. Drivers choose their travel route according to maximum personal utility, which is commonly assumed as reducing travel time and length, and minimizing cost. However, there are many secondary factors that significantly affect route planning. The inaccuracy derived from traditional methods for travel data collection (e.g. travel diary and retroactive questionnaires) becomes a limitation on reliability and makes driver behavior analysis more difficult to perform. Questioning drivers can be a tedious process both for the driver and the surveyor, and frequently, drivers do not completely succeed to recall the chosen route.

The presented research investigates route choice based on existing methods. It integrates behavioral latent factors evaluated by a stated preference survey and uses GPS as a measure for collecting and storing travel data in a more reliable way. The primary goal of this research is to develop a route choice model, which combines main latent factors that affect driver's behavior with quantitative service variables such as travel time and route length.

Two important goals are derived from the main goal and the problem definition:

  1. Using GPS receivers for collecting and storing travel data and analyzing them by GIS software including urban network travel data, in order to test the reliability and quality of data collected by combining GPS and GIS.
  2. Analyzing main latent factors that may affect driver's behavior while choosing routes, and integrating them with route choice models.

The research method was divided into three stages:

  • Performing a structured survey of stated preference in order to identify and evaluate behavioral latent factors that affect driver's decision when choosing routes.
  • Performing a travel survey in which 60 drivers performed a single daily trip that was recorded by a GPS receiver.
  • Developing a route choice model - at the final stage, a set of data was generated based on both survey analyses by which some probabilistic route choice models were tested (MNL , PSL, and NL).

The research contribution is a development of an efficient and reliable route choice model that will be based on GPS and GIS combined system for collecting, storing and analyzing travel data, and utilize traffic planners for better efficient design of future traffic network and improvement of existing networks.