CMF - 2015 Traffic Records Forum

Report
FUTURE CMF RESEARCH AND
CHALLENGES
Traffic Records Forum
October 27, 2014
Daniel Carter, UNC HSRC
Motivation
How do site characteristics and
countermeasures affect road
safety?
 CMF
How do site characteristics and
countermeasures affect a CMF?
 The question at hand
Research Project
• NCHRP 17-63 – Guidance on the Development and
Application of Crash Modification Factors
• UNC Highway Safety Research Center
• Vanesse Hangin Brustlin, Inc.
• Persaud and Lyon, Inc.
• Kittleson and Associates, Inc.
• Structured by three main objectives
OBJECTIVE 1.
TRANSFERABILITY
Develop guidelines for calibration of current CMFs to
assess treatment effectiveness at sites for which key
site characteristics may be different.
Transferability
Rural Kansas
Rural Washington
Rural Maine
Rural Connecticut
Questions
• When a countermeasure is applied to a different type of
•
•
•
•
site, will it be MORE effective?
Or LESS?
Or EQUAL?
When will this be the case?
What site characteristics will impact the CMF value?
Using Existing Data
• Information from the CMF Clearinghouse
• Treatments with many studies
• Cumulative meta-analysis
• CMFunctions
• Detailed disaggregate data analysis
• Information on individual locations
• Location-specific characteristics
Exploration of Influential Factors
Treatment
Influential
Factors on the
CMF
Area and
Facility
Type
Study Citation
Installation of
safety edge
treatment
Paved vs unpaved
shoulder types
Rural 2 lane
roads
Signal to
roundabout
conversion
Number of lanes (1
vs. 2), area type
(urban vs.
suburban), number
of intersection legs
(3 vs. 4)
Signalized
intersections
in urban and
suburban
areas
Graham, J.L., Richard, K.R. , O'Laughlin, M.K.,
Harwood, D.W., "Safety Evaluation of the Safety
Edge Treatment" Report No. FHWA-HRT-11-024,
Federal Highway Administration, Washington, DC.
(2011)
Srinivasan, R., Baek, J., Smith, S., Sundstrom, C.,
Carter, D., Lyon, C., Persaud, B., Gross, F.,
Eccles, K., Hamidi, A., and Lefler, N., "NCHRP
Report 705: Evaluation of Safety Strategies at
Signalized Intersections.", Washington, D.C.,
Transportation Research Board, National Research
Council, (2011)
Guidance to Produce
• Guidance on overall procedure for selecting and applying
a CMF
• Guidance on identifying influential factors for a CMF
• Preset list of influential factors for common safety treatments and
treatment categories
• Guidance for identifying influential factors if CMF is not in the
preset list
• Guidance on testing CMFs for homogeneity
• Guidance on adjusting CMFs for differences in site
conditions
• Physical characteristics
• Volume distribution
• Crash type and severity
OBJECTIVE 2. MULTIPLE
COUNTERMEASURES
Develop guidelines for how existing and future CMFs
can be combined in a single location with multiple
treatments.
Multiple Countermeasures
Questions
• Will multiple countermeasures have an effect that’s
GREATER than the individual countermeasures?
• Or LESS?
• Or EQUAL?
• And when will this be true?
Identified Existing Methods for Combining
CMFs
• Multiplicative Methods
• Dominant Effect Method
• Dominant Effect by Crash Type
• Dominant Common Residuals Method
• Estimating Combined Effect for Unknown Interaction
Define Ground Truth
Treatment Q
Level
1 (Not Present)
2 (Present)
A (Not Present)
A1
A2
B (Present)
B1
B2
Treatment P
Assess and Validate Methods
• Assess validity of identified methods (compare results of
methods to ground truth)
• Identify most possible scenarios of interaction effects
• Select best method (or develop a new one) and validate
against real world dataset
Guidance to Produce
• Guidance on combining single CMFs to estimate
combined effects
• Step-by-step procedures and examples for selecting and
applying appropriate method(s)
Example of Guidance to Produce
Define Applicability of Individual CMFs
(i.e., to what crash types and severities do the individual CMFs apply?)
Same Crash Type and Severity
Different Crash Type and Severity
Define Categories for Individual
Treatments
(e.g., roadway, roadside, intersection)
Define Categories for Individual
Treatments
(e.g., roadway, roadside, intersection)
Same General
Category
Different General
Category
Same General
Category
Different General
Category
Large Potential
Interaction Effect:
Apply Method 1
Medium Potential
Interaction Effect:
Apply Method 2
Small Potential
Interaction Effect:
Apply Method 3
Negligible Potential
Interaction Effect:
Apply Method 4
OBJECTIVE 3. CRASH
MODIFICATION FUNCTION
DEVELOPMENT
Develop recommended procedures for formulating and
calibrating future CMFs that identify key influential site
characteristics.
CMFunction Development
Questions
• How can CMFunctions be developed from regression
model coefficients?
• Key is to explore interactive effects between the CMF and non-
CMF related variables (e.g., interaction between centerline rumble
strips and AADT).
• How can CMFunctions be developed from CMF point
estimates?
• CMF point estimates from the same study with a variety of
application circumstances.
• CMF point estimates from different studies
• Combination of the first two scenarios
Guidance to Produce
• Guidance for developing CMFunctions and assessing fit
• Guidance for quantifying the CMF standard error when
computed as a function.
• Guidance on planning and designing studies (e.g.,
experimental) and treatments to better facilitate the
development of CMFunctions
Potential Data Sources
• Two-way to multi-way stop control (NC)
• Improved curve delineation (CT, WA)
• Stop to signal control conversion (NC, CA)
• Improving pavement friction (MN, PA)
Thank You
Daniel Carter
UNC Highway Safety Research Center
[email protected]

similar documents