laying on its side, roof, or even returning upright on all four wheels. Rollovers occur in a
multitude of ways. The risk of rollover is greater for vehicles designed with a high center of
gravity in relation to the track width. Driver behavior and road conditions are significant factors
in rollover crash events. Specifically, the factors that strongly relate to rollover fatalities are: if it
was a single-vehicle crash, if it was a rural crash location, if it was a high-speed roadway, if it
occurred at night, if there was an off-road tripping/tipping mechanism, if it was a young driver, if
the driver was male, if it was alcohol-related, if it was speed-related, if there was an unbelted
occupant, and if an occupant was ejected.
Rollover is one of the most severe crash types for light vehicles. In 2012, 112,000
rollovers occurred as the first harmful event, measuring 2 percent of the 5,615,000 police-
reported crashes involving all types of motor vehicles. In 2012, single, light-vehicle rollovers
accounted for 6,763 occupant deaths. This represented 20 percent of motor vehicle fatalities in
2012, 31 percent of people who died in light-vehicle crashes, and 46 percent of people who died
in light-vehicle single-vehicle crashes.
NHTSA describes rollovers as “tripped” or “untripped.” In a tripped rollover, the vehicle
rolls over after leaving the roadway due to striking a curb, soft shoulder, guard rail or other
object that “trips” it. Crash data suggest approximately 95 percent of rollovers in single-vehicle
crashes are tripped.
A small percentage of rollover events are untripped, typically induced by
tire and/or road interface friction. Whether or not a vehicle rolls when it encounters a tripping
mechanism is highly dependent upon the ratio of two vehicle geometric properties, referred to as
DOT HS 812 016, available at www-nrd.nhtsa.dot.gov/Pubs/812016.pdf.
See 68 FR 59251. Docket No. NHTSA-2001-9663, Notice 3. Available at https://federalregister.gov/a/03-25360
the Static Stability Factor (SSF). The SSF of a vehicle is calculated as one-half the track width, t,
divided by the height of the center of gravity (c.g.) above the road, h; SSF = (t/2h). The inertial
force that causes a vehicle to sway on its suspension (and roll over in extreme cases) in response
to cornering, rapid steering reversals or striking a tripping mechanism, like a curb or the soft
shoulder of the road, when the vehicle is sliding laterally, may be thought of as a force acting at
the c.g. to pull the vehicle body laterally. A reduction in c.g. height increases the lateral inertial
force necessary to cause rollover by reducing its leverage, and this is represented by an increase
in the computed value of SSF. A wider track width also increases the lateral force necessary to
cause rollover by increasing the leverage of the vehicle’s weight in resisting rollover, and that
advantage also increases the computed value of SSF. The factor of two in the computation (t/2h)
makes SSF equal to the lateral acceleration at which rollover begins in the most simplified
rollover analysis of a vehicle, which is represented by a rigid body without suspension
movement or tire deflections.
In 2001, the agency decided to use SSF to indicate rollover risk in a single-vehicle
Additionally, in that notice, the agency introduced the rollover resistance rating as a
means to quantify the risk of a rollover if a single-vehicle crash occurs. The agency emphasizes
that this rating does not predict the likelihood of a rollover crash occurring only that of a rollover
occurring given that a single vehicle crash occurs. In this rating system, the lowest rated vehicles
(1 star) are at least 4 times more likely to rollover than the highest rated vehicles (5 stars).
The rollover rating that was included as part of NCAP was based on a regression analysis
that estimated the relationship between single-vehicle rollover crashes and the vehicles’ SSF
For further explanation see the description and Figure 1 at
See 66 FR 3388. Docket No. NHTSA-2000-8298. Available at https://federalregister.gov/a/01-973.
C# HTML5 PDF Viewer SDK deployment on IIS in .NET
the place where you store XDoc.PDF.HTML5 Viewer Bit Applications" in accordance with the selected DLL (x86 site configured in IIS has no sufficient authority break pdf; break a pdf into separate pages
using state crash data. The SSF is measured at a Vehicle Inertial Measurement Facility
NHTSA acquires vehicles and measures the height of the vehicle c.g. The VIMF
consistently measures the c.g. height location of a particular vehicle using the stable pendulum
configuration. The test facility must be capable of measuring the c.g. height location to within
0.5 percent of the theoretical height, typically the 3-dimensional computer generated solid model
value of that vehicle. The track width is also measured on the same vehicle at this time. The risk
of rollover originally calculated for the 2001 notice was based on a linear regression analysis of
220,000 single-vehicle crash events reported by 8 States (Florida, Maryland, Missouri, New
Mexico, North Carolina, Ohio, Pennsylvania, and Utah).
Pursuant to the FY 2001 DOT Appropriations Act, NHTSA funded a National Academy
of Science (NAS) study on vehicle rollover resistance ratings.
The study focused on two
topics: whether the SSF is a scientifically valid measurement that presents practical, useful
information to the public, and a comparison of the SSF versus a test with rollover metrics based
on dynamic driving conditions that may include rollover events. NAS published their report at
the end of February 2002.
The NAS study found that SSF is a scientifically valid measure of rollover resistance for
which the underlying physics and real-word crash data are consistent with the conclusions that an
increase in SSF reduces the likelihood of rollover. It also found that dynamic tests should
complement static measures, such as SSF, rather than replace them in consumer information on
rollover resistance. The NAS study also made recommendations concerning the statistical
“The design of a Vehicle Inertial Measurement Facility,” Heydinger, G. J. et al, SAE Paper 950309, February,
Department of Transportation and Related Agencies Appropriations, 2001. Pub. L. 106-346 (Oct. 23, 2000).
“Rating System for Rollover Resistance, An Assessment,” Transportation Research Board Special Report 265,
National Research Council.
analysis of rollover risk and the representation of ratings methodology. The two primary
recommendations suggested using logistic regression rather than linear regression for analysis of
the relationship between rollover and SSF, and a high-resolution representation of the
relationship between rollover and SSF than is provided in the current 5-star program.
On October 14, 2003, NHTSA published a final policy statement outlining its changes to
the NCAP rollover resistance rating.
Beginning with the 2004 model year, NHTSA combined
a vehicle’s SSF measurement with its performance in a dynamic “fishhook” test maneuver
presented as a single rating. The fishhook maneuver is performed on a smooth pavement and is a
rapid steering input followed by an over-correction representative of a general loss-of-control
situation. This action attempts to simulate steering maneuvers that a driver acting in panic might
use in an effort to regain lane position after dropping two wheels off the roadway onto the
Additionally, the predicted rollover resistance ratings were reevaluated. Consistent with
the NAS recommendations, the agency changed from a linear regression to a logistic regression
analysis of the data. The sample size increased to 293,000 single-vehicle crash events, producing
a narrow confidence interval on the repeatability of the relationship between SSF and rollover. In
contrast, the linear regression analysis performed on the rollover rate of 100 make/models in
each of the six States providing data, resulted in a sample size of 600. In addition, a second risk
curve was generated for vehicles that experienced a tip-up in the dynamic fishhook test.
ii. Updates to the rollover NCAP SSF risk curve
Commenters to NHTSA’s 2008 NCAP upgrade notice asked NHTSA to collect crash
data on vehicles equipped with ESC in order to develop a new rollover risk model. In July 2008,
See 68 FR 59250. Docket No. NHTSA-2001-9663, Notice 3. Available at https://federalregister.gov/a/03-25360.
the agency upgraded the NCAP program to combine the rollover rating with the frontal and side
crash ratings, creating a single, overall vehicle rating.
No changes were made to the risk model
at that time.
However, NHTSA received comments requesting that the agency collect this
crash data to develop a new rollover risk model that better describes the rollover risk of all
vehicles that reflects the real-world benefits of ESC.
To enhance its rollover program, the
agency responded that they would continue to monitor the rollover rate for single-vehicle crashes
involving ESC equipped vehicles.
The accumulation of crash data involving vehicles equipped with ESC has been slow.
The 2003 regression analysis was based on 293,000 crash events. Up until recently, the agency
had observed fewer than 10,000 crashes with ESC-equipped vehicles. Previously, NHTSA was
not confident that it could accurately redraw the risk curves using such a small sample size. The
agency now believes that it has accumulated enough data to see a narrower tolerance band
adequate for use in a rating system.
According to the 2013 FARS, 7,500 vehicle occupants were killed in light-vehicle
These 2013 rollovers accounted for 34.6 percent of the 21,667 fatalities in light
vehicles that year. Of these 7,500 fatalities, 6,254 were killed in single-vehicle rollovers. NCAP
provides a consumer information rating program articulating the risk of rollover, to encourage
consumers to purchase vehicles with a predicted lower risk of a rollover. This information
enables prospective purchasers to make choices about new vehicles based on differences in
rollover risk and serve as a market incentive to manufacturers to design their vehicles with
See 73 FR 40021. Docket No. NHTSA-2006-26555. Available at https://federalregister.gov/a/E8-15620.
See 73 FR 40032. Docket No. NHTSA-2006-26555. Available at https://federalregister.gov/a/E8-15620.
See 72 FR 3475. Docket No. NHTSA-2006-26555. Available at https://federalregister.gov/a/E7-1130.
Traffic Safety Facts 2012. DOT HS 812 032 available at www-nrd.nhtsa.dot.gov/Pubs/812032.pdf.
greater rollover resistance. The consumer information program also informs drivers, especially
those who choose vehicles with poorer rollover resistance, that their risk of harm can be greatly
reduced with seat belt use to avoid ejection. The program seeks to remind consumers that even
the highest rated vehicle can roll over, but that they can reduce their chance of being killed in a
rollover by about 75 percent just by wearing their seat belts.
NHTSA intends to update and recalculate the risk curve using ESC data collected from
20 States, and to transition the rollover risk rating into a new crash avoidance rating. In this new
rollover scoring, NHTSA would not be changing the dynamic rollover test. The agency believes
that embedding rollover into the crash avoidance rating is more appropriate since it targets
rollover prevention and it also consolidates the message of reduced crash incidence. Rollover
resistance would remain a significant component in the rating scheme, weighted based on its
relative importance to overall vehicle safety. The details of how the crashworthiness rating is
combined with the crash avoidance rating into an overall rating system are discussed in the rating
section of this RFC notice.
The statistical model created in 2003 combined SSF and dynamic maneuver test
information to predict rollover risk. The agency performed the Fishhook test on about 25 of the
100 make/model vehicles for which SSF was measured and substantial State crash data was
Eleven of the 25 vehicles tipped up
in the Fishhook maneuver that was conducted
in the heavy condition with a 5-occupant load. All 11 vehicles had SSFs less than 1.20.
An Experimental Examination of 26 Light Vehicles Using Test Maneuvers That May Induce On-Road,
Untripped Light Vehicle Rollover—Phase VI of NHTSA’s Light Vehicle Rollover Research
Program, NHTSA Technical Report, DOT HS 809 547, 2003.
A “tip-up” occurs when the two vehicle wheels lift off the ground 2 inches during the Fishhook test.
At that time, the agency believed it was very unlikely that passenger cars would tip-up in
the maneuver test because no tip-ups were observed in the passenger cars tested at the low end of
the SSF range for passenger cars. To validate that assumption, the agency tested a few passenger
cars each year at the low end of the SSF range. No tip-ups have been observed in the agency tests
for any vehicle type since 2007. Therefore, the agency is unable to produce an estimate or a
logistic regression curve based on tip/no-tip as a variable.
The rollover statistical model was populated with new data and used logistic regression
analysis to update the rollover risk curve. The agency examined 20 State datasets for single-
vehicle crashes involving vehicles equipped with ESC that occurred during 2011 and 2012. Data
were reported by Delaware, Florida, Iowa, Illinois, Indiana, Kansas, Kentucky, Maryland,
Michigan, Missouri, Nebraska, New Jersey, New Mexico, New York, North Carolina, North
Dakota, Pennsylvania, Washington, Wisconsin, and Wyoming. The dataset was comprised of
11,647 single-vehicle crashes, of which 627 resulted in rollover. For 2011, NHTSA used data
reported by each of the 20 States for single-vehicle crashes involving ESC-equipped vehicles; a
summation of 5,429 crashes. For 2012, NHTSA used data reported by 10 States for single-
vehicle crashes involving ESC-equipped vehicles; 6,218 crashes. Table 8 shows a summary of
the 2011 and 2012 State dataset used for the logistic regression analysis.
Table 8. Summary of 2011 and 2012 State data used to generate the rollover risk curve
The new dataset included 197 different makes/models for which the SSF had been
calculated within NCAP; the SSF ranged from 1.07 to 1.53. The new dataset contained two
vehicle types, passenger cars and light truck vehicles, including pickup trucks, SUVs, and vans.
To accomplish the rollover analysis, it is more appropriate to use the state dataset because it
provides the ability to filter for ESC-equipped vehicles rather than the NHTSA FARS database,
which is not sufficiently granular. FARS contains two data elements; rollover and rollover
location. The rollover data element has attributes of no rollover, tripped rollover, untripped
rollover, and unknown type rollover. The rollover location data element has attributes of no
rollover, on roadway, on shoulder, on median/separator, in gore, on roadside, outside of
trafficway, in parking lane/zone, and unknown. The State dataset distribution compares similarly
to the FARS number of vehicles involved in fatal crashes with a rollover occurrence. Table 9
summarizes the 2011 and 2012 rollover data for the number of single-vehicle crashes for ESC-
equipped vehicles by vehicle type. For comparison, Table 10 summarizes the number of vehicles
involved in fatal crashes with a rollover occurrence by vehicle type, as reported in FARS. In the
new rollover model dataset, pickup trucks appear to be slightly underrepresented and SUVs
appear to be slightly overrepresented compared with the FARS data.
Table 9. Summary of 2011 and 2012 State data used to generate the rollover risk curve
Source: State Data System
Table 10. Vehicles involved in fatal crashes with a rollover occurrence
2011 + 2012
The agency performed a logistic regression analysis of the 11,647 single-vehicle crash
events. The dependent variable in this analysis is vehicle rollover, while the independent
variables are SSF, light condition, driver age, driver gender, and the State indicator variable. The
SAS® logistic regression program used these variables to compute the model. The SAS®
statistical analysis software output tables are available in the docket for this RFC notice. Figure 4
shows a plot of the predicted rollover probability versus the SSF for the 20-State dataset. Figure
5 is a plot of the average predicted probability of rollover for each SSF in the dataset. Figures 4
and 5 demonstrate the relationship between SSF and the predicted probability of rollover, that at
every level of SSF the predicted probability of rollover is less than it was estimated to be in
2003. The flatter curve for the 2011 + 2012 dataset aligns with increased vehicle SSFs, the
expected effect of ESC on rollover frequency, and the reduced observation of rollover in single-
Figure 4. Current and new rollover risk curve
Documents you may be interested
Documents you may be interested