Loading...

Urban Accident Prediction Models based on Urban Transportation Planning Process

Mohseni Ashtiani, Elnaz | 2013

540 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 44417 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Nassiri, Habibollah
  7. Abstract:
  8. In recent years, most of the researches in Urban Transportation Planning have been in the areas of prediction and management of transportation demand, increasing the capacity of roads and reducing traffic congestion. In spite of several studies which were performed in the field of collisions prediction modeling for highways and freeways, these models were not applicable to long-term transportation planning process due to their inability to accurately predict variables for future long-term transportation planning.In this study, Aggregate Crash Prediction Models (ACPMs) were performed using the Socio-economics, land use and street network information of 253 traffic zone. This information was used for the Comprehensive Transportation Planning study of City of Mashhad, Iran. In addition, five types of available GIS collisions information (i.e., property-damage-only, injury, fatal, severe injury+fatal, and total) for the same traffic zones were used. Results of the STATA statistical software show that the regression models in a log-link form, and a negative binomial (NB) error structure provided the best fits. Aerial distance from each zone to the Haram (i.e., zone 1) and the commercial area were found to be the two significant variables in these models.Also, of the model developed, the total number accidents model was found to be the most accurate one. Models of property-damage-only, severe accidents and injury accidents were in the subsequent order of the accuracy, respectively.
  9. Keywords:
  10. Transportation Planning ; Macroscopic Cpllision Prediction ; Generalized Linear Model ; Traffic Analysis Zone

 Digital Object List

 Bookmark

No TOC