2006•Transportation Research Board 85th Annual MeetingTransportation Research BoardRequires access

Exploring Relationship Between Safety and Consistency of Geometry and Speed on British Roads

Roger N. Bird, Ibrahim Hassan Hashim

Open publisher page 5 citations

Abstract

Rural single carriageways (two lane highways) present a particular challenge to highway engineers as they constitute the majority of the network in terms of distance (kilometres), but the total distance traveled on them is low (measured in vehicle kilometres). The number of accidents per kilometer tends to be low, but the rate of accidents per vehicle kilometer is high. Highway consistency studies aim to give guidance as to how road improvements can be targeted to make the best use of available resources. A large sample of rural roads from the north-east of England was analyzed in this study. Over 1200 curves and tangents were identified on 380km of derestricted road. A wide variety of alignment and speed indices were defined and calculated for each element. Accident records for these roads over a five year period were obtained and analyzed. Accidents were subdivided by severity and by number of vehicles involved. Using generalized Linear modeling, a negative binomial or Poisson model was created for each index individually (along with traffic flow and section length) against each of the five accident groups. Combinations of the significant indices are then examined to define the best predictive model for each accident group. It was found that there is a strong correlation between accident frequencies and various consistency measures. Different measures fitted better to different accident subgroups, and element type. A general model was derived that could be applied to all elements for all personal injury accidents.

About this research paper

What this paper is about

Rural single carriageways (two lane highways) present a particular challenge to highway engineers as they constitute the majority of the network in terms of distance (kilometres), but the total distance traveled on them is low (measured in vehicle kilometres). The number of accidents per kilometer tends to be low, but the rate of accidents per vehicle kilometer is high. Highway consistency studies aim to give guidance as to how road improvements can be targeted to make the best use of available resources. A large sample of rural roads from the north-east of England was analyzed in this study. Over 1200 curves and tangents were identified on 380km of derestricted road. A wide variety of alignment and speed indices were defined and calculated for each element. Accident records for these roads over a five year period were obtained and analyzed. Accidents were subdivided by severity and by number of vehicles involved. Using generalized Linear modeling, a negative binomial or Poisson model was created for each index individually (along with traffic flow and section length) against each of the five accident groups. Combinations of the significant indices are then examined to define the best predictive model for each accident group. It was found that there is a strong correlation between accident frequencies and various consistency measures. Different measures fitted better to different accident subgroups, and element type. A general model was derived that could be applied to all elements for all personal injury accidents.

Why it matters

OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Rural single carriageways (two lane highways) present a particular challenge to highway engineers as they constitute the majority of the network in terms of distance (kilometres), but the total distance traveled on them is low (measured in vehicle kilometres). The number of accidents per kilometer tends to be low, but the rate of accidents per vehicle kilometer is high. Highway consistency studies aim to give guidance as to how road improvements can be targeted to make the best use of available resources. A large sample of rural roads from the north-east of England was analyzed in this study. Over 1200 curves and tangents were identified on 380km of derestricted road. A wide variety of alignment and speed indices were defined and calculated for each element. Accident records for these roads over a five year period were obtained and analyzed. Accidents were subdivided by severity and by number of vehicles involved. Using generalized Linear modeling, a negative binomial or Poisson model was created for each index individually (along with traffic flow and section length) against each of the five accident groups. Combinations of the significant indices are then examined to define the best predictive model for each accident group. It was found that there is a strong correlation between accident frequencies and various consistency measures. Different measures fitted better to different accident subgroups, and element type. A general model was derived that could be applied to all elements for all personal injury accidents.

Key concepts: Kilometer, Consistency (knowledge bases), Poisson regression, Poisson distribution, Negative binomial distribution, Geometric design, Transport engineering, Statistics

Related papers

Back to paper searchBrowse research topicsOriginal source
Exploring Relationship Between Safety and Consistency of Geometry and Speed on British Roads — Research Paper | ScholarLens