Accident prediction models for unsignalized intersections
The main objective of this thesis is to develop Accident Prediction Models (APM) for estimating the safety potential of urban unsignalized (T and 4-leg) intersections in the Greater Vancouver Regional District (GVRD) and Vancouver Island on the basis of their traffic characteristics. The models a...
Main Author: | |
---|---|
Format: | Others |
Language: | English |
Published: |
2009
|
Online Access: | http://hdl.handle.net/2429/7882 |
id |
ndltd-UBC-oai-circle.library.ubc.ca-2429-7882 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-UBC-oai-circle.library.ubc.ca-2429-78822018-01-05T17:33:54Z Accident prediction models for unsignalized intersections Rodríguez, Luis F. (Luis Felipe) The main objective of this thesis is to develop Accident Prediction Models (APM) for estimating the safety potential of urban unsignalized (T and 4-leg) intersections in the Greater Vancouver Regional District (GVRD) and Vancouver Island on the basis of their traffic characteristics. The models are developed using the generalized linear regression modeling (GLIM) approach, which addresses and overcomes the shortcomings associated with the conventional linear regression approach. The safety predictions obtained from GLIM models can be refined using the Empirical Bayes' approach to provide, more accurate, site-specific safety estimates. The use of the complementary Empirical Bayes approach can significantly reduce the regression to the mean bias that is inherent in observed accident counts. The thesis made use of sample accident and traffic volume data corresponding to unsignalized (both T and 4-leg) intersections located in urban areas of the Greater Vancouver Regional District (GVRD) and Vancouver Island. The data included a total of 427 intersections located in the cities of Victoria, Surrey, Nanaimo, Coquitlam, Burnaby and Vancouver. The information available for each intersection included the total number of accidents in the 1993-1995 period, traffic volumes for both major and minor roads given in Average Annual Daily Traffic (AADT) and type of intersection (T or 4-leg). Four categories of models were developed in this study: (1) models for the total number of accidents; (2) separate models for T and 4-leg intersections; (3) separate models for different regions (Vancouver Island, the Lower Mainland and Surrey); and (4) a model for Surrey including intersection control. Five applications of APM were used in this thesis. Four of them relate to the use of the Empirical Bayes refinement: identification of accident-prone locations, developing critical accident frequency curves, ranking the identified accident-prone locations and before and after safety evaluation. The fifth application provides a safety-planning example, comparing the safety of a 4-leg intersection to two staggered T-intersections. These applications show the importance of implementing APM as a tool to assess in a reliable fashion traffic safety, and design different safety strategies. Applied Science, Faculty of Civil Engineering, Department of Graduate 2009-05-05T17:11:04Z 2009-05-05T17:11:04Z 1998 1998-05 Text Thesis/Dissertation http://hdl.handle.net/2429/7882 eng For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use. 4113956 bytes application/pdf |
collection |
NDLTD |
language |
English |
format |
Others
|
sources |
NDLTD |
description |
The main objective of this thesis is to develop Accident Prediction Models (APM) for estimating
the safety potential of urban unsignalized (T and 4-leg) intersections in the Greater Vancouver
Regional District (GVRD) and Vancouver Island on the basis of their traffic characteristics. The
models are developed using the generalized linear regression modeling (GLIM) approach, which
addresses and overcomes the shortcomings associated with the conventional linear regression
approach. The safety predictions obtained from GLIM models can be refined using the Empirical
Bayes' approach to provide, more accurate, site-specific safety estimates. The use of the
complementary Empirical Bayes approach can significantly reduce the regression to the mean bias
that is inherent in observed accident counts.
The thesis made use of sample accident and traffic volume data corresponding to unsignalized
(both T and 4-leg) intersections located in urban areas of the Greater Vancouver Regional
District (GVRD) and Vancouver Island. The data included a total of 427 intersections located in
the cities of Victoria, Surrey, Nanaimo, Coquitlam, Burnaby and Vancouver. The information
available for each intersection included the total number of accidents in the 1993-1995 period,
traffic volumes for both major and minor roads given in Average Annual Daily Traffic (AADT)
and type of intersection (T or 4-leg). Four categories of models were developed in this study: (1)
models for the total number of accidents; (2) separate models for T and 4-leg intersections; (3)
separate models for different regions (Vancouver Island, the Lower Mainland and Surrey); and
(4) a model for Surrey including intersection control.
Five applications of APM were used in this thesis. Four of them relate to the use of the Empirical
Bayes refinement: identification of accident-prone locations, developing critical accident
frequency curves, ranking the identified accident-prone locations and before and after safety
evaluation. The fifth application provides a safety-planning example, comparing the safety of a
4-leg intersection to two staggered T-intersections. These applications show the importance of
implementing APM as a tool to assess in a reliable fashion traffic safety, and design different
safety strategies. === Applied Science, Faculty of === Civil Engineering, Department of === Graduate |
author |
Rodríguez, Luis F. (Luis Felipe) |
spellingShingle |
Rodríguez, Luis F. (Luis Felipe) Accident prediction models for unsignalized intersections |
author_facet |
Rodríguez, Luis F. (Luis Felipe) |
author_sort |
Rodríguez, Luis F. (Luis Felipe) |
title |
Accident prediction models for unsignalized intersections |
title_short |
Accident prediction models for unsignalized intersections |
title_full |
Accident prediction models for unsignalized intersections |
title_fullStr |
Accident prediction models for unsignalized intersections |
title_full_unstemmed |
Accident prediction models for unsignalized intersections |
title_sort |
accident prediction models for unsignalized intersections |
publishDate |
2009 |
url |
http://hdl.handle.net/2429/7882 |
work_keys_str_mv |
AT rodriguezluisfluisfelipe accidentpredictionmodelsforunsignalizedintersections |
_version_ |
1718587806963466240 |