Development and Application of Crash Severity Models for Highway Safety: Conduct of Research Report (2023) / Chapter Skim
Currently Skimming:
Pages 4-25
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 4... ...
4 2 LITERATURE REVIEW This chapter provides a thorough literature review of previous research on methodologies that have been used for estimating crash counts by injury severity. The following sections report on the literature we reviewed in the following areas: 1.
|
From page 5... ...
5 for the impact of exogenous variables to vary across the count alternatives. In addition to observed factors, the multivariate models inherently account for correlation across multiple crash frequency variables for an observation unit.
|
From page 6... ...
6 et al.
|
From page 7... ...
7 Table 2.1 Summary of Existing Aggregate Level Crash Severity Studies Studies Spatial Unit Region Number of Levels Explored Methodological Approach Independent Variables Considered Roadway Infrastruc- ture Land- use Built Environment Traffic Characteristics Sociodemo- graphic Wea- ther Count Framework (Aguero-Valverde and Jovanis, 2006) County level, (macro)
|
From page 8... ...
8 Studies Spatial Unit Region Number of Levels Explored Methodological Approach Independent Variables Considered Roadway Infrastruc- ture Land- use Built Environment Traffic Characteristics Sociodemo- graphic Wea- ther (Aguero-Valverde and Jovanis, 2009) Statemaintained rural twolane roads (micro)
|
From page 9... ...
9 Studies Spatial Unit Region Number of Levels Explored Methodological Approach Independent Variables Considered Roadway Infrastruc- ture Land- use Built Environment Traffic Characteristics Sociodemo- graphic Wea- ther (Narayanamoorthy et al., 2013) Census tract (macro)
|
From page 10... ...
10 Studies Spatial Unit Region Number of Levels Explored Methodological Approach Independent Variables Considered Roadway Infrastruc- ture Land- use Built Environment Traffic Characteristics Sociodemo- graphic Wea- ther (Zhan et al., 2015) Census tract (Macro)
|
From page 11... ...
11 Studies Spatial Unit Region Number of Levels Explored Methodological Approach Independent Variables Considered Roadway Infrastruc- ture Land- use Built Environment Traffic Characteristics Sociodemo- graphic Wea- ther (Boulieri et al., 2017) Ward (macro)
|
From page 12... ...
12 Studies Spatial Unit Region Number of Levels Explored Methodological Approach Independent Variables Considered Roadway Infrastruc- ture Land- use Built Environment Traffic Characteristics Sociodemo- graphic Wea- ther Fractional Split Framework (Milton et al., 2008) Highway segments (micro)
|
From page 13... ...
13 2.2 COMPARISON OF EXISTING METHODS FOR ESTIMATING CRASH COUNTS BY SEVERITY LEVELS To better understand the differences in the existing methods for estimating crash counts by severity, in this section, three methods for estimating crash counts by severity levels, namely, the Highway Safety Manual (HSM) method (1st Edition)
|
From page 14... ...
14 where: N = predicted average crash frequency on the roadway segment for base conditions, AADT = annual average daily traffic volumes (vehicles/day) on the roadway segment, L = length of the roadway segment (miles)
|
From page 15... ...
15 Strengths The advantage of the HSM crash severity model is simplicity. Model calculations are easy.
|
From page 16... ...
16 The value of the overdispersion parameter associated with Y is determined as a function of length for twolane and multi-lane rural facility segments as follows: ππ = 1 exp[ππ+ln(πΏπΏ)
|
From page 17... ...
17 β’ By subtracting KA from KAB, one can only get the prediction for B type of crash but not the SPF for B type, in particular no over-dispersion can be easily calculated. 2.2.3 Negative Binomial-Ordered Probit Fractional Split Models Summary The approach β referred to as the Negative Binomial-Ordered Probit Fractional Split (NB-OPFS)
|
From page 19... ...
19 π»π»ππππ in our model takes the ordered probit probability (Ξ) form for the severity category ππ.
|
From page 20... ...
20 Model Assumption The approach employs a simple mathematical conversion of crash counts by severity to crash proportions by severity. The non-linear conversion results in an entirely different functional form relative to the multivariate models.
|
From page 21... ...
21 Table 2.3 Methodologies Used for Crash Severity Modeling Methodological Approach Previous Research Artificial Neural Networks Delen et al.
|
From page 22... ...
22 2.4 FACTORS AFFECTING CRASH INJURY SEVERITY OUTCOMES Factors that affect the results of crash injury severities upon the occurrence of a crash have been widely investigated by previous research. As the focus of this project is modeling frameworks that can better predict crash counts by severity, it is critical to understand what factors were found by previous research to significantly affect crash injury severity outcomes that are not always considered in ordinary crash count models.
|
From page 23... ...
23 Table 2.4 Summary of Significant Variables in Affecting Injury Severity Outcomes Variable Names Previous Literature Driver Age Al-Ghamdi, 2002; Rifaat & Tay, (2009) ; Kockelman & Kweon, 2002; Islam & Mannering, (2006)
|
From page 24... ...
24 2.5 CATEGORIZATION OF DRIVERS BASED ON DEMOGRAPHICS AND VEHICLE TYPES A significant difference between the methods implemented in this project and the previous crash severity count models is that here we incorporate driver and vehicle information into the models. Driver and vehicle information in conventional safety data is available only at the "disaggregate" level, that is, disaggregated by crash.
|
From page 25... ...
25 therefore result in high severity. Moreover, in single-vehicle crashes, pickups and sport utility vehicles have greater severities than passenger cars.
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.