Heavy Truck Stability

UNIVERSITY OF MICHIGAN TRANSPORTATION RESEARCH INSTITUTE• OCTOBER–DECEMBER 2000 • VOLUME 31, NUMBER 4 •

Rollover of HeavyCommercial Vehicles

ISSN 0739 7100October–December 2000, Vol. 31, No. 4

Editor: Monica MillaDesigner and Illustrator: Shekinah ErringtonOriginal Technical Illustrations: Chris WinklerPrinter: UM Printing Services

The UMTRI Research Review is published four times a year bythe Research Information and Publications Center of the Uni-versity of Michigan Transportation Research Institute, 2901Baxter Road, Ann Arbor, Michigan 48109-2150 (http://www.umtri.umich.edu). The subscription price is $35 a year,payable by all subscribers except those who are staff membersof a State of Michigan agency or an organization sponsoringresearch at the Institute. See the subscription form on the insideback cover. For change of address or deletion, please encloseyour address label.

The University of Michigan, as an equal opportunity/affirmative action em-ployer, complies with all applicable federal and state laws regarding nondis-crimination and affirmative action, including Title IX of the Education Amend-ments of 1972 and Section 504 of the Rehabilitation Act of 1973. The Univer-sity of Michigan is committed to a policy of nondiscrimination and equal op-portunity for all persons regardless of race, sex, color, religion, creed, nationalorigin or ancestry, age, marital status, sexual orientation, disability, or Vietnam-era veteran status in employment, educational programs and activities, or ad-missions. Inquiries or complaints may be addressed to the University's Directorof Affirmative Action and Title IX/Section 504 Coordinator, 4005 WolverineTower, Ann Arbor, Michigan 48109-1281, (734) 763-0235, TDD (734) 647-1388. For other University of Michigan information call (734) 764-1817.

The Regents of the University:David A. Brandon, Ann ArborLaurence B. Deitch, Bloomfield HillsDaniel D. Horning, Grand HavenOlivia P. Maynard, GoodrichRebecca McGowan, Ann ArborAndrea Fischer Newman, Ann ArborS. Martin Taylor, Gross Pointe FarmsKatherine E. White, Ann ArborLee Bollinger, President, ex officio

UMTRI RESEARCH REVIEW 1

Rollover of HeavyCommercial Vehicles

ROLLOVER ACCIDENTS ANDVEHICLE ROLL STABILITY

Rollover accidents of commer-

cial vehicles are especially violent

and cause greater damage and injury

than other accidents. The relatively

low roll stability of commercial

trucks promotes rollover and

contributes to the number of

truck accidents.

Rollover and Accident SeverityThere are over 15,000 rollovers of commer-

cial trucks each year in the U.S. That is about onefor every million miles of truck travel. About 9,400of these—about one for every four million miles—are rollovers of tractor-semitrailers.

Commercial truck rollover is strongly associ-ated with severe injury and fatalities in highwayaccidents. As shown in figure 1, about 4 percentof all truck accidents involve rollover, but morethan 12 percent of fatalities in truck accidents in-volve rollover (General Estimates System, 1995and Truck and Bus Crash Fact Book, 1995).

The association of rollover with injuries to thetruck driver is even stronger. While only 4.4 per-cent of tractor-semitrailer accidents are rollovers,58 percent of the fatal injuries to the truck driveroccurred in rollover crashes (General EstimatesSystem and Trucks Involved In Fatal Accidents,1992–1996). Figure 1 shows that rollover is over-represented in all forms of truck-driver injury andthat the level of overrepresentation increases pro-gressively with the severity of injury.

Roll Stability and the Occurrence ofRollover AccidentsThe low level of basic roll stability of com-

mercial trucks sets them apart from light vehiclesand appears to be a significant contributing causeof truck rollover accidents. The basic measure ofroll stability is the static rollover threshold, ex-pressed as lateral acceleration in gravitational units(g). Most passenger cars have rollover thresholds

by Chris Winkler

2 OCTOBER—DECEMBER 2000

greater than 1 g, while light trucks, vans, and SUVsrange from 0.8 to 1.2 g (Chrstos, 1991 and Tech-nical Assessment Paper: Relationship BetweenRollover and Vehicle Factors, 1991). However, therollover threshold of a loaded heavy truck oftenlies well below 0.5 g.

The typical U.S. five-axle tractor-van semi-trailer combination, when loaded to legal grossweight, has a rollover threshold as high as 0.5 gwith an optimal high-density, low center-of-grav-ity (cg) load. This drops to as low as 0.25 g with aworst-case load that completely fills the volumeof the trailer (Ervin et al., 1980 and 1983). Thetypical U.S. five-axle petroleum semitanker has arollover threshold of about 0.35 g (Ervin andMathew, 1988). Rollover thresholds of commoncryogenic tankers that transport liquefied gasesare as low as 0.26 g. El-Gindy and Woodrooffefound a variety of logging trucks operating inCanada with thresholds ranging from 0.23 to 0.31g(Ervin and Nisonger, 1982). Individual vehicleswith rollover thresholds well below 0.2 g alsooccur occasionally (e.g., Sweatman, 1993).

Drivers maneuver their vehicles at well over0.2 g fairly regularly. The AASHTO guidelinesfor highway curve design result in lateral accel-erations as high as 0.17 g at the advised speed (APolicy on Geometric Design of Highways andStreets, 1990). Therefore, even a small degree ofspeeding beyond the advisory level will easily

cause actual lateral accelerations to reach 0.25 gin everyday driving. On the other hand, tire fric-tional properties limit lateral acceleration on flatroad surfaces to a bit under 1 g at the very most.These observations clearly imply that the rolloverthreshold of light vehicles lies above, or just mar-ginally at, the extreme limit of the vehicle’smaneuvering ability, but the rollover threshold ofloaded heavy trucks extends well into the “emer-gency” maneuvering capability of the vehicle andsometimes into the “normal” maneuvering range.

Nevertheless, it is relatively hard for truck driv-ers to perceive their proximity to rollover whiledriving. Rollover is very much an either/or situa-tion. It is something like walking up to a cliff withyour eyes closed: As you approach the edge, youare still walking on solid ground but once you’vestepped over, it’s too late. Further, the rolloverthreshold of a commercial truck changes regularlyas the load changes, so drivers may not have thechance to get used to the stability of their vehicle.Finally, especially for combinations, the flexiblenature of the tractor frame tends to isolate thedriver from the roll motions of the trailer, whichmight act as a cue to rollover. These observationssuggest the following safety hypotheses:

• Heavy trucks are more susceptible thanlight vehicles to rollover accidents causeddirectly by inadvertently operating the

Figure 1. Rollover is strongly associated with accident severity and with serious injury to truck drivers.


Figure 2. The rollover threshold of trucksextends deep into the maneuvering range.

Figure 3. Untripped rollovers are common fortractor-semitrailer combinations but rare for cars.

vehicle beyondthe rolloverthreshold.

• Rollover inheavy-truckaccidents isstrongly re-lated to thebasic roll sta-bility of thevehicle.

The first hy-pothesis describessingle-vehicle ac-cidents in whichthe first signi-ficant event isan untripped roll-over. Unfortunate-ly, the perfect ac-cident file for suchan analysis does not exist, but GES files for 1993–1996 (which do not indicate first event) show thatuntripped rollovers occur in more than 20 percentof single-vehicle rollover accidents for tractorsemitrailers, but in less than 4 percent of thoseaccidents for passenger cars. Further, the Trucksin Fatal Accidents files (which contain no compa-rable data for cars) for 1994–1996 show thatuntripped, first-event rollovers account for 26.8percent of single-vehicle rollover accidents.

Between 79 and 84 percent of single-vehiclerollover crashes on highway ramps are first-eventuntripped rollovers in which the vehicle struck noother object prior to rolling over (Wang and Coun-cil, 1999 and Council and Chen, 1999).

Figure 4 shows a strong relationship betweenphysical roll stability and the chance of rolloverin a single-vehicle accident. The relationship isnonlinear; that is, as the vehicle becomes moreand more stable, the chance of rollover asymp-totically approaches zero. Conversely, asstability decreases, the sensitivity of the probabil-ity of rollover to stability increases rapidly andthe function becomes quite steep.

Figure 4 demonstrates that, as roll stability de-clines to low levels, the probability of rollover inan accident increases rapidly until the vehiclebecomes very likely to roll over in nearly any


accident. Moreover, for the low-stability vehiclesfor which rollover is such a great concern, rela-tively small improvements in physical stability canyield rather large improvements in rollover acci-dent rate.

THE MECHANICS OFSTATIC ROLL STABILITY

All rollover events in the real world aredynamic events; none are truly quasi-static. How-ever, the foregoing analyses of accident data showa very strong relationship between the basic, staticroll stability of the heavy vehicle and the actualoccurrence of rollover in accidents. Accordingly,this section considers the mechanics of quasi-staticrollover to show how this fundamental performance

property derives from the mechanical behaviorof the various components of the vehicle.

Figure 5 presents a simplified model of a heavyvehicle in a steady turn in which the vehicle, itstires, and suspensions have been “lumped” into asingle roll plane. The nomenclature of the figureis as follows:

ay

is lateral acceleration

Fi

are the vertical tire loads, i=1, 2

h is the height of the cg

T is the track width

W is the weight of the vehicle

∆y is the lateral motion of the cg relative tothe track

φ is the roll angle of the vehicle

Figure 4. The chance of rollover is strongly influenced by the roll stabilityof the vehicle.


The equilibrium equation for roll momentabout a point on the ground at the center of thetrack is:

W•h•ay = (F

2 – F

1)•T/2 – W•∆y

Qualitatively, two destabilizing (overturning)moments act on the vehicle:

• A moment due to the lateral D’Alambertforce acting through the cg, W•h•a

y , as a

result of the external imposition of lateralacceleration

• A moment due to the weight of the ve-hicle acting at a position that is laterallyoffset from the center of the track, W•∆y

The first moment results from the externalimposition of lateral acceleration acting at thecenter of gravity (cg) of the vehicle; the latterresults from the internal compliant reactions ofthe vehicle (roll and lateral shift).

These two destabilizing moments are op-posed by one stabilizing (restoring) moment dueto the side-to-side transfer of vertical load onthe tires, (F

2 - F

1)•T/2 . This moment is also due

to the internal, compliant responses of thevehicle. The maximum possible value of this mo-ment is W•T/2, which occurs when all load is

transferred to one sideof the vehicle, i.e.,when F

2 = W and F

1 = 0.

One way of inter-preting the equationand the observation oftwo destabilizing mo-ments is that a vehicle’srollover threshold de-rives from both areference rigid-bodystability, which wouldresult if ∆y were zero,and the degradationfrom that reference re-sulting from the lateralmotion of the cg al-lowed by complianceswithin the vehicle.

Figure 6 illustrateshow various propertiesof the vehicle contrib-ute to the rolloverthreshold according tothis view. The exampleshows a rather low-stability vehicle whoseheavy load and rela-tively high payloadestablish a rigid-bodystability of 0.45 g.However, the roll andlateral motions allowedby the various compli-ances and free-plays inthe tires,suspension, chassisstructures, and even theload itself, can reducethe actual static stabil-ity of the vehicle toabout 0.26 g.

Starting from thetop of the figure, if thisvehicle were rigid, thenas a turn became moresevere and lateral ac-celeration increased, it could transfer all of its loadonto the outside tires without suffering any lat-eral shift of the cg (∆y in figure 5). This means itFigure 5. A simplified freebody diagram of a

heavy vehicle in a steady turn.

Figure 6. An example case showing variousmajor influences that determine roll stability.


could achieve the maximum stabilizing moment(for load transfer) without suffering anydestablizing moment from lateral shift. For a givencg height and track, the vehicle would be optimized.

But real vehicles are not rigid. As load is trans-ferred from the tires toward the inside of the turnonto the tires toward the outside, those tires de-flect causing roll motion about the center of thetrack. As a result, the cg moves outboard and somestability is lost.

Similarly, the suspension springs deflect. Theroll from this motion takes place about the

so-called roll center of the suspension, typicallylocated a few inches above or below the axle.

Some suspension, and virtually all tractor-to-trailer couplings, are designed with some freeplay

that comes into play only whenthe vehicle rolls substantially.This freeplay allows an incre-ment of roll motion to take placewithout any attendant stabilizinglateral load transfer. Theresulting lateral shift of the cgfurther degrades stability.

The structural elements of thevehicle can simply bend underthe high centrifugal loads thatdevelop during severe

cornering. Further, the payload itself may de-flect sideways as it suffers under the same type ofloading. Figure 7 illustrates some of the more sig-nificant of these deflections. Figure 8 shows—inrather dramatic fashion—the torsional complianceof the vehicle’s structural frame can also contrib-ute to the rollover process.

The significance of lateral displacements oc-curring from each individual mechanism such asthese—or of the total of all—can be judged bycomparing ∆y to T/2 (the half track). The lateraldisplacement of the cg is, in effect, a direct reduc-

tion of the half track. Inround numbers, thehalf track of an axlewith dual tires is about95 cm. Thus, a 1-cmlateral deflection re-sults in loss of stabil-ity equal to about 1percent of the originalrigid-body stability ofthe vehicle. Lateralsuspension deflectionmay be on the order of2 cm. Lateral beamingof the trailer may be 3cm or more. A varietyof other compliancesmay each produce dis-placements on theorder of several milli-meters and, of course,the lateral offset of theplacement of the cargo

Figure 7. Examples of other mechanisms that can contribute to thedestabilizing offset moment.

Figure 8. The rear end of a torsionally compliant flat-bed trailer rolls over nearlyindependently of the front end.


can be quite substantial. While none of these dis-placements may seem significant individually, thetotal influence can easily account for the loss of asignificant portion of the rigid-vehicle stability.

The overall message of figure 6, and this dis-cussion, is that roll stability is established by thesummated effects of many compliance mecha-nisms. While the effect of any one compliance maybe small, virtually all compliances degrade sta-bility. All the compliances combined can reduceroll stability to as little as 60 percent of the ideal-ized rigid-vehicle stability.

Measuring Static Rollover Threshold withthe Tilt-Table ExperimentThe tilt-table test provides a highly resolute

method of determining rollover threshold and aconvenient means for examining the mechanismsby which this limit is determined. The methodologyis a physical simulation of the roll-plane experienceof a vehicle during quasi steady-state turning.

In this experimental method, the vehicle isplaced on a tilt table and is very gradually tiltedin roll. As shown in figure 9, the component ofgravitational forces parallel to the table surfaceprovides a simulation of the centrifugal forcesexperienced by a vehicle in turning maneuvers.The progressive application of these forcesachieved by slowly tilting the table serves to

simulate the effects of quasi-statically increasinglateral acceleration in progressively more severe,steady turnings. The tilting process continues un-til the vehicle reaches the point of roll instabilityand “rolls over.” (The vehicle is constrained bysafety straps to prevent actual rollover.)

When the table is tilted, the component ofgravitational forces parallel to the table surfacesimulates lateral forces, and the weight of thevehicle is simulated by the component of gravita-tional forces perpendicular to the table. Both thelateral and vertical forces acting during the tilt-table test are scaled down somewhat relative tothe real forces they simulate. The amount thatthese forces are scaled depends on the amount oftilt required. This scaling has multiple effectswhich, although they tend to cancel one another,can nevertheless reduce the accuracy of the ex-periment. The quality of the result as a measureof the true static stability limit of the vehicle de-pends, in part, on not requiring an excessively largetilt angle to achieve rollover. Because heavy ve-hicles are relatively unstable, they typicallydo not require a large tilt angle, and thereforethe experiment is very well suited for examiningthese vehicles.

The fundamental aspects of the mechanics ofquasi-static rollover, which were discussed brieflyabove, have been confirmed in numerous tilt-tableexperiments (Winkler, 1987; Winkler and Zhang,1995; and Ervin et al., 1998).

Figure 9. The tilt-table experiment.


DYNAMIC CONSIDERATIONS IN HEAVYVEHICLE ROLLOVERS

Analyses of the accident records make it clearthat static roll stability is the dominant vehiclequality affecting the chance of a given heavy truckbeing involved in a rollover accident. The previ-ous section reviewed the mechanics of staticstability. However, all rollover accidents in the realworld are dynamic events to some extent; noneare truly quasi-static. This section examines someinfluences of dynamics on rollover.

Simple Dynamics in the Roll PlaneQuasi-static rollover is nearly impossible to ac-

complish, even on the test track. The analyses inthe previous sectionassume that the lateralacceleration conditionis a given and is sus-tained indefinitely(i.e., the condition de-fining steady state).

In practice, a testvehicle can approachrollover quasi-stati-cally either by veryslowly increasing turnradius at a constantvelocity or by veryslowly increasing ve-locity at a constantradius. In either case,the quasi-static condi-tion can be made tohold reasonably welluntil the tires of thedrive axles lift. At this

point, however, the vehicle typically loses trac-tion and “scrubs off” speed such that the lateralacceleration immediately declines and the drivewheels settle back onto the surface. The processmay be repeated any number of times. At leasttwo exceptions can allow quasi-static rollover: Thevehicle may be equipped with a locking differentialso that drive thrust can be maintained after lift oftires on the drive axles, or highly compliant (flatbed) trailers may roll over at the rear withoutlifting drive-axle tires (figure 8). Regardless, inreal-world events there is virtually always a

dynamic component to the maneuver which, atthe least, provides the needed kinetic energy toraise the cg through its apex height after the tiresof all axles (or at least all axles other than the steeraxle) have left the ground. However, as shown infigure 10, for vehicles with high centers of grav-ity, the additional elevation of the cg required isnot that great.

Several simplified analyses describing mini-mum requirements for dynamic rollover (i.e., asdepicted in figure 10) exist in the literature. Thesetend to focus on the passenger car and, conse-quently, on so-called tripped rollovers, i.e.,rollovers involving a curb-strike or other mecha-nism that may produce lateral tire forces well inexcess of those generally obtainable on a flat, hardroad surface (e.g., Rice et al.).

Cooperrider et al. take a different approach.They present an analysis based on a constantlateral tire force applied to a rigid vehicle over asustained period of time. This approach seemsmore applicable to rollover of commercial ve-hicles, particularly in situations of sustained,quasi-steady turning. Cooperrider’s results showthat the lateral acceleration needed to producerollover is a function of the length of time it isapplied. If the acceleration can be sustained in-definitely, it need only equal the static stabilitylimit (T/2h for this rigid-vehicle analysis). But ifthe lateral acceleration exceeds the static limit, itneed only be sustained for a finite time to resultin rollover. For example, for a typical heavy truck,acceleration of 110 percent of the static limit canproduce rollover if sustained for about 1 second;120 percent need be sustained for only about0.6 seconds.

Dynamic Considerationsin Transient ManeuversDynamics become particularly important when

the frequency content of the maneuver (and inparticular, the lateral acceleration that results frommaneuvering) approaches or exceeds the naturalfrequency of the rolling motion of the vehicle. Alightly loaded tractor-semitrailer can be expectedto have natural frequencies in roll in the range of2 Hz or more—well above the frequency of steer-ing input that the truck driver can muster even inemergency maneuvers. However, a heavily loadedvehicle, with its payload cg at a moderate height

Figure 10. At the least, rollover requires thedynamic momentum required to lift the cgthrough its apex height.


and with suspensions of average roll stiffness, islikely to exhibit a roll natural frequency near 1 Hz.A heavily loaded semitrailer with a high cg andwith suspensions of less-than-average stiffness canhave a roll natural frequency as low as 0.5 Hz. Asindicated below, 0.5 Hz in particular is well withinthe range of excitation frequencies expected inemergency maneuvering. Thus, one can expect thepotential for harmonic tuning and related resonantovershoot to promote rollover in transient maneu-vers with higher frequency content. It follows fromthese considerations that high levels of roll stiff-ness (and consequently roll natural frequency) andof roll damping generally promote dynamic roll sta-bility in highway operations.

Higher frequency maneuvers also involve yawdynamics that can complicate—and stabilize—roll behavior of articulated vehicles. Figure 11shows the response of a tractor-semitrailer duringa simulated, 2-second emergency lane change ma-neuver (ISO 14791). The figure presents time

Figure 11. In a dynamic maneuver, the acceleration of the semitrailerlags the tractor and roll lags acceleration.

histories of lateral acceleration for the tractor andfor the semitrailer and roll angle for the combina-tion. When maneuvering at speed, the semitrailertends to follow the path of the tractor rather faith-fully. Particularly with longer vehicles, this impliesa time lag between the actions of the tractor andthe trailer. (This is more a result of the tractrixgeometry that basically governs the motion of thetrailer, rather than a true dynamic phenomenon.)When the frequency content of the lateral motionapproaches the roll natural frequency, roll motioncan be expected to lag lateral acceleration. Bothof these effects are readily apparent in figure 11.With respect to rollover, when the trailer reachesits maximum roll displacement, the tractor is wellpast its peak lateral acceleration. Consequently,at this critical point, the tractor, with its relativelylow cg, is more “available” to resist rollover thanit would be in a demanding steady-state turn. Thus,in this maneuver, while roll dynamics are degradingroll stability, the yaw dynamics are compensatingto some extent. The situation (even in this rela-tively simple maneuver) is complex and the netresult depends on the tuning of the frequency con-tent of the particular maneuver, the frequencysensitivities of the vehicle in yaw, and the naturalfrequency and damping of the vehicle in roll.

Dynamics can play a unique role in therollover of multiply-articulated vehicles. Asillustrated in figure 12, vehicles with more

Figure 12. In rapid obstacle-avoidance maneu-vers, rearward amplification may result inpremature rollover of the rear trailer.


Numerous approaches to reduce rearwardamplification of multitrailer vehicles have beenproposed, most of which are based on differentarrangements for coupling trailers. The most suc-cessful have been the so-called B-train andC-train, which are compared to the referenceA-train in figure 13. Both of these vehicles elimi-nate the yaw and roll degrees of freedom associatedwith the pintle-hitch coupling between the semi-trailer and the full trailer. Eliminating the yawarticulation indirectly improves roll stability byreducing rearward amplification. For example, theA-train in figure 13 would typically have a rear-ward amplification of about 2, but the rearwardamplification of the B-train and C-train in the fig-ure would typically be less than 1.5.

However, by coupling the two trailers in roll,the B- and C-train configurations dramaticallyimprove dynamic roll stability. The lateral accel-eration and roll motions of the two trailers areabout 90 degrees out of phase. Thus, when thesecond trailer reaches its critical condition ofmaximum lateral acceleration and roll angle, thefirst trailer has passed its peak and returned tonear-zero in these two measures and actually hassubstantial roll momentum in the opposite direc-tion. When these two trailers are coupled in rollas in a B- or C-train, the vehicle can perform verysevere lane changes (i.e., with peak lateral accel-erations of the tractor on the order of 0.5 g)without experiencing rollover because it is ex-tremely difficult for one trailer to “drag over” itsout-of-phase partner (Winkler et al., 1986).(Of course, the mechanical loads on the couplerand dolly frame may be very high in such maneu-vers, introducing the risk of mechanical failureof these parts.)

than one yaw-articulation joint (e.g., truck-trailercombinations, doubles, or triples) may exhibit anexaggerated response of the rearward units whenperforming maneuvers with unusually high fre-quency content. The phenomenon is known asrearward amplification and is often quantified, asshown in figure 12, by the ratio of the peak lateralresponse of the rearward unit to that of the tractor(ISO 614791 and SAE J2179).

Rearward amplification is a strong function ofthe frequency content (and the type) of the ma-neuver. Because rearward amplification is closeto unity at low frequencies, these vehicles behavevery well in normal driving. However, since rear-ward amplification tends to peak in the frequencyrange characteristic of quick, evasive maneuvers,these vehicles are also quite susceptible to rolloverof the rear trailers during emergency maneuvering.

Figure 13. The B-train and C-train, originally introduced in Canada,exhibit less rearward amplification than the standard A-train.


and Krupka, 1985). Fundamental analyses ofsloshing liquids in road tankers appeared in theliterature from the 1970s. A number of moreelaborate computer studies arose in the late1980s and early 1990s. This discussion is con-strained to basic elements that provide insighton the mechanisms by which fluid motions in-fluence rollover. The mechanisms of slosh aremost readily described in simple steady-statecornering, although it is in transient maneuversthat the most exaggerated fluid displacementstake place.

Steady TurningWhen a slosh-loaded tanker performs a

steady-state turn, the liquid responds to lateralacceleration by displacing laterally, keeping itsfree surface perpendicular to the combinedforces of gravity and lateral acceleration. Fig-ure 14a illustrates the position of a partialliquid load in a circular tank subjected to asteady-state cornering maneuver. The mass cen-ter of the liquid moves on an arc, the center ofwhich is at the center of the circular tank. Ineffect, the shift of the liquid produces forceson the vehicle as if the mass of the load werelocated at the center of the tank.

THE INFLUENCE OF SLOSHING LIQUIDS

In the majority of commercial truck operations,the load on the vehicle is fixed and nominally cen-tered. In certain cases, however, the load may beable to move in the vehicle, with the potential ofaffecting the turning and rollover performance.The most common examples of moving loads arebulk, liquid tankers with partially filled compart-ments; refrigerated vans hauling suspended meatcarcasses; and livestock. The performance prop-erties of commercialvehicles used in theseapplications may beinfluenced by the freemovement of the loadin either longitudinalor lateral directions.This section presentsmaterial on the firsttwo types of loads.

Liquid LoadsThe most important

of these is liquid cargocarried in tanks. In theoperation of a bulk-liquidtransport vehicle, themoving load that canaffect its cornering androllover behavior is thepresence of unre-strained liquid due toa partially filled tank orcompartment. A com-partment that is filled to anything less than its fullcapacity allows the liquid to move from side toside, producing a “slosh” load condition. Slosh isof potential safety concern because the lateral shiftof the load reduces the vehicle’s performance incornering and rollover, and the dynamic motionsof the load may occur out of phase with thevehicle’s lateral motions in such a way as to be-come exaggerated and thus further reduce therollover threshold.

The motions of liquids in a tank vehicle can bequite complex due to the dependence of the mo-tions on tank size and geometry, the mass andviscosity of the moving liquid, and the maneuverbeing performed (Dalzell, 1967; Komatsu, 1987;

Figure 14. Liquid position in steady-state turning for circular and rectangular tanks.


With more complex tank shapes, even thesteady-state behavior becomes somewhat difficultto analyze. In particular, with unusual tank shapesit becomes more difficult to describe the motionof the liquid’s center of mass as a function of lat-eral acceleration. As a contrast to the circular tank,figure 14b illustrates the behavior of liquid in arectangular tank. At low lateral accelerations, theliquid movement is primarily lateral, centered ata point well above the tank center. Hence, its ef-fect is similar to having a very high mass center.With increasing lateral acceleration, the mass cen-ter follows a somewhat elliptical path.

While the circular tank results in a vehicle witha higher load center, efforts to reduce the loadheight by widening and flattening the tank can beexpected to increase vehicle sensitivity to sloshdegradation of the rollover threshold. The effectis illustrated by the plot in figure 15 which isadapted from data by Strandberg (Ranganathanlater presented very similar results). The figureshows rollover threshold versus load condition insteady-state cornering. For a circular tank, increas-ing load lowers the threshold continuously due tothe increasing mass of fluid free to move sideways.In this case, the minimum rollover threshold oc-curs at full load. For a vehicle with a modifiedrectangular tank, higher levels of rollover thresh-old occur when the tank is either empty or full,although at intermediate load conditions therollover threshold is severely depressed due to thegreater degree of lateral motion possible for theunrestrained liquid. Thus, the rectangular tankshape (in contrast to the circular) can potentiallyresult in rollover thresholds with sloshing loadsthat are less than that of the fully loaded vehicle.

Transient TurningIn transient maneuvers such as an abrupt eva-

sive steering maneuver (e.g., a rapid lane change),slosh loads introduce the added dimension of dy-namic effects. With a sudden steering input, therapid imposition of lateral acceleration may causethe fluid to displace to one side with anunderdamped (overshooting) type of behavior. Thedifference between the steady-state and transientmaneuvers is primarily a matter of the time in-volved in entering the turn. The steady-state typeof behavior is observed when the turn is enteredvery slowly, whereas the transient behavior appliesto a very rapid turning maneuver. The response

Figure 16. Rollover threshold in a transientturn as a function of the percentage of loadof unrestrained liquid (adapted fromStrandberg, 1978).

Figure 15. Rollover threshold in a steadyturn as a function of the percentage of loadof unrestrained liquid and tank shape(adapted from Strandberg, 1978).


of the liquid mass to a step input of accelerationwould be seen to displace to an amplitude approxi-mately twice the level of the steady-state amplitude.In a lane-change maneuver in which the accelera-tion goes first in one direction and then the other,an even more exaggerated response amplitude canbe produced.

In general, the degree to which the dynamicmode is excited depends on the timing of themaneuver. The unrestrained liquid will have anatural frequency for its lateral oscillation whichdepends on the liquid level and cross-sectional sizeof the tank. For a half-filled, eight-foot-widetanker, this frequency is approximately 0.5 Hz(cycles per second), while a six-foot-diameter cir-cular tank (typical of an 8,800-gallon tanker)would have a frequency of approximately 0.6 Hz.As for dynamic systems in general, if the fre-quency content of input (lateral acceleration) staysbelow this natural frequency, the response islargely quasi-static, but if the input contains sub-stantial power at or above the natural frequency,the response will be dynamic. Although they donot do so in normal driving, drivers in emergencysituations are generally capable of generatingsteering inputs at frequencies in the range of 0.5 Hz(e.g., McLean and Hoffmann, 1973). Indeed, thetwo-second lane change used as a typical evasivemaneuver for evaluating rearward amplification con-stitutes a lateral acceleration input at just thatfrequency closely matched to the slosh frequency.Hence it must be concluded that dynamic sloshmotions can be readily excited on a tanker of normalsize, especially in the course of evasive maneu-vers such as a lane change.

In transient maneuvers, rollover thresholds aredepressed by this dynamic motion. Figure 16 showsthe estimated rollover threshold as a function ofload for unrestrained liquids in a transientmaneuver, which is adapted from data presentedby Strandberg. In the transient case, even the cir-cular tank experiences reduced rollover thresholdswhen partially loaded because the fluid can “over-shoot” the steady-state level. Understandably, theelliptical tanker is even worse. Though the resultsshown are derived from analytical studies, experi-mental tests of partially loaded tankers generallyconfirm these observations (Culley et al., 1978).

Partial Liquid LoadsIn the vocational use of many liquid bulk haul-

ers, it is sometimes necessary to run with partialloads. This is especially true with local deliverytankers hauling gasoline and home-heating fuel.The question is: What can be done to reduce thesensitivity and hence the potential risks of usingthese vehicles, once a substantial fraction of theirload has been delivered? Of course, specifying avehicle with suspension systems most resistantto rollover is a first step. However, at least twoother aids are available:

Baffles. Baffles are commonly used in tankvehicles, except in special cases where provisionsfor cleaning prevent their use (such as bulk-milkhaulers). However, the common baffle arrange-ment is a transverse baffle intended to impedefore/aft movement of the load. These transversebaffles have virtually no utility in preventing thelateral slosh influential to roll stability. To improveroll performance, longitudinal baffles would berequired, but design and cost considerations havepractically eliminated their use.

Compartmentalization. A more commonmethod for improving cornering performancewith tankers under partial loading conditions isto subdivide the tank into separate compartments.Ideally, the compartments are completely emp-tied on an individual basis at a drop spot so thevehicle is never subject to a sloshing load. Theonly precaution in this type of use is that the de-livery route should be planned to empty from therear of the vehicle first. When it is not possible tocompletely empty each compartment, a reducedslosh sensitivity exists, but is often not signifi-cant as long as only a fraction of the total load isfree to slosh. In these cases, the relevant param-eters are the percent of load being carried and thefraction of the load that is free to slosh.


ROLLOVER AND THE INTELLIGENTHIGHWAY/VEHICLE SYSTEM

Modern electronics are beginning to be appliedto the problem of heavy vehicle rollover in theform of intelligent systems in the vehicle or withinthe highway infrastructure.

Because a disproportionate number (about7 percent) of commercial-vehicle rollover crashesoccurs on ramps, highway-infrastructure systemshave concentrated on active signing for advisoryspeeds on exit ramps. The methods vary signifi-cantly. For example, Freedman et al. examinedthe effectiveness of speed-advisory signs withflashing lights that activated when a truck wasobserved entering the ramp at excessive speed.On the other hand, Strickland et al. described pro-totype installations that selectively display themessage “Trucks reduce speed,” based on auto-mated observations of the speed, weight, andheight of individual vehicles. Systems of this typehave been installed and monitored at three differentexit ramps on the Capital Beltway in Washing-ton, D.C. Prior to their installation, truck rolloversoccurred once every year or every other year onthese ramps. After installation, there were no truckrollovers at any of the sites for the three-year periodof the study (Strickland and McGee, 1998).

At least three methods of reducing commer-cial-vehicle rollover through on-board systems are

being pursued. Perhaps the most direct method isactive roll control, which aims to improve the rollstability of vehicles during critical events.Kusahara et al. describe a prototype active rollstabilizer installed on the front suspension of amedium duty commercial truck. Similar devices,installed on all suspensions of unit trucks or trac-tor-semitrailer combinations, have been underdevelopment at Cambridge University (Lin et al.,1994 and 1996).

Another approach employing on-board intelli-gence is the roll-stability-advisory (RSA) orrollover-warning system. A “stability monitoringand alarm system” was advertised for applicationon commercial vehicles in the late 1980s (Preston-Thomas and Woodrooffe, 1990). More recently,Roaduser Research of Melbourne, Australia, hasdeveloped and installed a rollover-warningsystem in limited numbers on tank vehicles. Thesystem produces an audible warning for the driverbased on real-time measurement of lateral accel-eration compared to a predetermined, worst-casestatic rollover threshold for the vehicle. UMTRI hasdeveloped a prototype RSA that includes a visualdisplay for the driver that compares the current lat-eral acceleration of the vehicle to the static rolloverthreshold of the vehicle in left- and right-hand turns.The rollover thresholds are calculated in real timebased on signals from on-board sensors. Thresh-olds for each new loading condition are determinedafter only a few minutes of normal driving.

Another approach to reducing rollover crashesis active yaw control of the vehicle, which preventslateral acceleration from exceeding the rolloverthreshold of the vehicle. The approach selectivelyapplies individual wheel brakes to submit appro-priate yaw moments and/or to simply slow thevehicle. Palkovics, in association with El-Gindy andothers, has published research articles on thisapproach and the ideas are being introduced in com-mercial applications. In addition, UMTRI hasdeveloped and demonstrated a prototype system es-pecially for reducing rearward amplification inmultitrailer vehicles (Ervin et al., 1998 and Winkleret al., 1998). Development of this system continueswith expectations of commercial application.

This article is a condensed version of an SAE publication.Copyright 2000 Society of Automotive Engineers, Inc. A com-plete copy of this report can be obtained from SAE by contactingSAE Customer Service at: 724-776-4970 (phone), 724-776-0790(fax), or [email protected], order number RR-004.

Figure 17. The driver display of the UMTRI RSA (Ervin et al., 1998).

estimated limitfor right-siderollover

estimated limitfor left-side

rollover

digital displayscurrent lateral acceleration

recall recent maximums

0.44 0.18 0.40


REFERENCES

A policy on geometric design of highways andstreets—1990; AASHTO green book.American Association of State Highway andTransportation Officials, Washington, D.C.

Chrstos, J. P. 1991. An evaluation of static rolloverpropensity measures. Interim final report.National Highway Traffic Safety Administra-tion, Vehicle Research and Test Center, EastLiberty, Ohio. Sponsor: National HighwayTraffic Safety Administration, Washington,D.C. Report No. VRTC-87-0086/ DOT/HS807 747.

Cooperrider, N.; Thomas, T.; Hammound, S. 1990.Testing and analysis of vehicle rolloverbehavior. Report No. SAE 900366.

Council, F. M.; Chen, L. 1999. Personal corre-spondence. HSIS data request. FederalHighway Administration. Washington, D. C.

Culley, C.; Anderson, R. L.; Wesson, L. E. 1978.Effect of cargo shifting on vehicle handling.Final report. Dynamic Science, Inc., Phoenix,Ariz. Sponsor: Federal Highway Admini-stration, Bureau of Motor Carrier Safety,Washington, D.C. Report No. 3989-78-22/FHWA-RD-78-76.

Dalzell, J. F. 1967. Exploratory studies of liquidbehavior in randomly excited tanks: lateral

excitation. Southwest Research Institute, SanAntonio, Texas. Sponsor: National Aero-nautics and Space Administration, George C.Marshall Space Flight Center, Huntsville,Ala. Report No. 2.

El-Gindy, M.; Woodrooffe, J. H. F. 1990. Study ofrollover threshold and directional stabilityof log hauling trucks. National ResearchCouncil Canada, Division of MechanicalEngineering, Ottawa, Ontario. Report No.TR-VDL-002/ NRCC No. 31274.

Ervin, R. D. 1983. The influence of size andweight variables on the roll stability of heavyduty trucks. Michigan University, Ann Arbor,Transportation Research Institute. Report No.SAE 831163.

Ervin, R. D.; Mallikarjunarao, C.; Gillespie, T. D.1980. Future configuration of tank vehicleshauling flammable liquids in Michigan.Volume I. Final report. Highway SafetyResearch Institute, Ann Arbor, Michigan.Sponsor: Michigan Department of State High-ways and Transportation, Lansing. Report No.UM-HSRI-80-73-1.

Ervin, R. D.; Mathew, A. 1988. Reducing the riskof spillage in the transportation of chemicalwastes by truck. Final report. Michigan Uni-versity, Ann Arbor, Transportation ResearchInstitute. Sponsor: Rohm and Haas, Bristol,Pa. Report No. UMTRI-88-28.

Chris Winkler is a research sci-entist in the Engineering ResearchDivision at UMTRI. He receiveda B.S. degree in mechanical engi-neering from Bucknell Universityand an M.S. in mechanical engi-neering from the University ofMichigan. He joined UMTRI in1969, and has been involved withthe analysis and prediction of the dynamicperformance of all pneumatic-tired vehicles,but with special emphasis on commercialhighway vehicles. Chris has been involvedwith both theoretical and experimentalwork on trucks, including the development

of specialized equipment for mea-suring the mechanical properties ofvehicles and their components.Recently, many UMTRI projects in-volve “intelligent vehicle systems,”and, in particular, activesystems to enhance stability. Recentprojects include development andreal-world testing of Roll Stability

Advisory systems. Chris is active with theSociety of Automotive Engineers and the In-ternational Standards Organization, and he isa trustee and the vice president for NorthAmerica of the International Forum for RoadTransport Technology.

ABOUT THE AUTHOR


Ervin, R. D.; Nisonger, R. L. 1982. Analysis of theroll stability of cryogenic tankers. Finalreport. Highway Safety Research Institute, AnnArbor, Michigan. Sponsor: Union CarbideCorporation, Linde Division, Tarrytown, N.Y.Report No. UM-HSRI-82-32.

Ervin, R. D.; Nisonger, R. L.; MacAdam, C. C.;Fancher, P. S. 1983. Influence of size andweight variables on the stability and controlproperties of heavy trucks. Volume I. Finalreport. Michigan University, Ann Arbor, Trans-portation Research Institute. Sponsor: FederalHighway Administration, Washington, D.C.Report No. FHWA-RD-83-029/ UMTRI-83-10/1.

Ervin, R.; Winkler, C.; Fancher, P.; Hagan, M.;Krishnaswami, V.; Zhang, H.; Bogard, S. 1998.Two active systems for enhancing dynamicstability in heavy truck operations. MichiganUniversity, Ann Arbor, Transportation ResearchInstitute. Sponsor: National Highway TrafficSafety Administration, Washington, D.C.Report No. UMTRI-98-39.

Freedman, M.; Olson, P. L.; Zador, P. L. 1992.Speed actuated rollover advisory signs fortrucks on highway exit ramps. Insurance Insti-tute for Highway Safety, Arlington, Virginia /Michigan University, Transportation ResearchInstitute, Ann Arbor.

ISO 14791. Road vehicles—Heavy commercialvehicle combinations and articulatedbuses—Lateral stability test procedures.International Organization for Standardization,Geneva, Switzerland.

Komatsu, K. 1987. “Non-linear sloshing analysis ofliquid in tanks with arbitrary geometries.”Japan National Aerospace Laboratories, Tokyo.International Journal of Nonlinear Mechanics,Vol. 22, No. 3, pp. 193–207.

Krupka, R. M. 1985. Mathematical simulation of thedynamics of a military tank. Northrop Corpor-ation, Electro-Mechanical Division, Anaheim,California. Report No. SAE 850416.

Kusahara, Y.; Li, X.; Hata, B.; Watanabe, Y. 1994.“Feasibility study of active roll stabilizer forreducing roll angle of an experimental medium-duty truck.” Nissan Diesel Motor Company,Ltd., Saitama, Japan. International Symposiumon Advanced Vehicle Control. Proceedings.Tokyo, Society of Automotive Engineers ofJapan, pp. 343-348. Report No. SAE 9438501.

Lin, R. C.; Cebon, D.; Cole, D. J. 1994. “Aninvestigation of active roll control of heavy roadvehicles.” Cambridge University, EngineeringDepartment, England. Shen, Z., ed. TheDynamics of Vehicles on Roads and on Tracks.Proceedings. 13th IAVSD Symposium. Lisse,Swets and Zeitlinger, 1994, pp. 308–321.

Lin, R. C.; Cebon, D.; Cole, D. J. 1996. “Active rollcontrol of articulated vehicles.” CambridgeUniversity, Department of Engineering,England. Vehicle System Dynamics, Vol. 26,No. 1, July 1996, pp. 17–43.

Lin, R. C.; Cebon, D.; Cole, D. J. 1996. “Optimalroll control of a single-unit lorry.” CambridgeUniversity, England. Institution of MechanicalEngineers. Proceedings. Part D: Journal ofAutomobile Engineering, Vol. 210, pp. 45–55.

McGee, H. W.; Strickland, R. R. 1994. “Anautomatic warning system to prevent truckrollover on curved ramps.” Bellomo-McGee,Inc., Vienna, Va. Public Roads, Vol. 57, No. 4,Spring 1994, pp. 17–22.

McLean, J. R.; Hoffmann, E. R. 1973. “The effectsof restricted preview on driver steering controland performance.” Melbourne University,Department of Mechanical Engineering,Australia. Human Factors, Vol. 15, No. 4,August 1973, pp. 421–430. Sponsor: AustralianRoad Research Board, Victoria.

Palkovics, L.; El-Gindy, M. 1993. “Examination ofdifferent control strategies of heavy-vehicleperformance.” National Research CouncilCanada, Ottawa, Ontario. Ahmadian, M., ed.Advanced Automotive Technologies—1993.Proceedings. New York, American Society ofMechanical Engineers, pp. 349–362.

Palkovics, L.; El-Gindy, M. 1995. “Design of anactive unilateral brake control system for five-axle tractor-semitrailer based on sensitivity.”National Research Council Canada, Ottawa,Ontario. Vehicle System Dynamics, Vol. 24, No.10, December 1995, pp. 725–758.

Palkovics, L.; Michelberger, P.; Bokor, J.; Gaspar, P.1996. “Adaptive identification for heavy-truckstability control.” Budapest TechnicalUniversity, Hungary. Segel, L., ed. TheDynamics of Vehicles on Roads and on Tracks.Proceedings. 14th IAVSD Symposium. Lisse,Swets and Zeitlinger, 1996, pp. 502–518.Sponsor: Hungarian Scientific Research andDevelopment Fund.


Palkovics, L.; Semsey, A.; Ilosvai, L.; Koefalvy, G.1994. “Safety problems and control of a typicalEuropean truck/tandem-trailer vehicleconfiguration.” Budapest Technical University,Hungary/ Hungarocamion Road SafetyDivision, Hungary. ISATA InternationalSymposium on Automotive Technology andAutomation, 27th. Proceedings for theDedicated Conference on Road and VehicleSafety. Croydon, Automotive Automation Ltd.,1994, pp. 303–310. Sponsor: HungarianNational Scientific Research Fund. Report No.94SF010.

Preston-Thomas, J.; Woodrooffe, J. H. F. 1990. Afeasibility study of a rollover warning devicefor heavy trucks. National Research CouncilCanada, Vehicle Dynamics Laboratory, Ottawa,Ontario. Sponsor: Transportation DevelopmentCentre of Canada, Montreal, Quebec. ReportNo. TP 10610E/TR-VDL-005.

Ranganathan, R.; Rakheja, S.; Sankar, S. 1989.“Kineto-static roll plane analysis of articulatedtank vehicles with arbitrary tank geometry.”Concordia University, Montreal, Quebec,Canada. International Journal of VehicleDesign, Vol. 10, No. 1, pp. 89–111. Sponsor:Canada, Transport Canada, TransportationDevelopment Centre, Montreal, Quebec.UMTRI-57246.

Rice, R. S.; Segal, D. J.; Jones, I. S. 1978.Development of vehicle rollover maneuver.Volume I. Final report. Calspan Corporation,Advanced Technology Center, Buffalo, NewYork. Sponsor: National Highway Traffic SafetyAdministration, Washington, D.C. Report No.CAL-ZQ-5993-V-1/ DOT/HS 804 093.

SAE J2179. 1993. A test for evaluating therearward amplification of multi-articulatedvehicles. Society of Automotive Engineers.Warrendale, Pennsylvania.

Strandberg, L. 1978. Lateral stability of roadtankers. Volume I—main report. Statens Vaeg-och Trafikinstitut, Linkoeping, Sweden.Sponsor: Statens Trafiksaekerhetsverk, Solna,Sweden; Transportforskningsdelegationen,Stockholm, Sweden. Report No. 138A.

Strickland, R.; McGee, H. 1998. Evaluation ofprototype automatic truck rollover warningsystems. Bellomo-McGee, Inc., Vienna,Virginia. Sponsor: Federal HighwayAdministration, Office of Safety and TrafficOperations, Washington, D.C. Report No.FHWA-RD-97-124.

Sweatman, P. F. 1993. Overview of dynamicperformance of the Australian heavy vehiclefleet. Road User Research Pty., Ltd.,Williamstown, Victoria, Australia. Sponsor:Australian National Road TransportCommission, Melbourne, Victoria. Report No.Technical Working Paper No. 7.

Technical assessment paper: relationship betweenrollover and vehicle factors. 1991. NationalHighway Traffic Safety Administration,Washington, D.C.

Truck and bus crash factbook. 1995. Special report.Michigan University, Ann Arbor,Transportation Research Institute. Sponsor:Office of Motor Carriers, Federal HighwayAdministration, Washington, D.C. Report No.UMTRI-97-30.

Wang, Jun; Council, Forest M. 1999. “Estimatingtruck rollover crashes on ramps using a multi-state database.” Appalachian RegionalCommission. University of North Carolina.Transportation Research Record. No. 1686.Wasshington D.C. Paper no. 99-0657, pp. 29–35.

Winkler, C. B. 1987. Experimental determination ofthe rollover threshold of four tractor-semitrailer combination vehicles. Final report.Michigan University, Ann Arbor, Trans-portation Research Institute. Sponsor: SandiaLaboratories, Albuquerque, N.M. Report No.UMTRI-87-31.

Winkler, C. B.; Fancher, P. S.; Carsten, O.; Mathew,A.; Dill, P. 1986. Improving the dynamicperformance of multitrailer vehicles: a study ofinnovative dollies. Volume I—technical report.Final report. Michigan University, Ann Arbor,Transportation Research Institute. Sponsor:Federal Highway Administration, Washington,D.C. Report No. UMTRI-86-26/I/ FHWA/RD-86/161.

Winkler, C. B.; Zhang, H. 1995. Roll stabilityanalysis of the TARVAN. Final report. MichiganUniversity, Ann Arbor, Transportation ResearchInstitute. Sponsor: Naval Air Warfare Center,Indianapolis, Ind. Report No. UMTRI-95-24.

Winkler, C., Fancher, P., Ervin, R. 1998. Intelligentsystems for aiding the truck driver in vehiclecontrol. The University of MichiganTransportation Research Institute. Report No.SAE 1999-01-1301. Warrendale, Pennsylvania.


UMTRI RESEARCH NOTES

Investigation of Hip Fractures and Aortic Injuriesin Motor Vehicle Crashes

NHTSA is sponsoring a five-year research program at UMTRI to investigate different types ofinjuries resulting from motor vehicle crashes with a specific focus on hip and aorta injuries. LarrySchneider, head of UMTRI’s Biosciences Division, is the project director and Chris Van Ee, assistantresearch scientist in the Biosciences Division, is the principal investigator. This program bringstogether both analysis of CIREN investigations of real-world automotive crashes and controlled bio-mechanical testing in the laboratory to define and investigate automotive injury mechanisms.

While many of the biomechanical factors of the knee and femur injuries have been investigated,the important factors leading to disabling hip injuries in frontal crashes remain largely unknown.Reducing the stiffness of the knee contact point (bolster/instrument panel) has been shown to decreaseth incidence of knee fractures. However, because of the lack of experimental data, it is unknown ifthese energy-absorbing knee bolsters are protecting the knees at the expense of the hip. Currently it isestimated that approximately 10,000 cases of hip fractures and dislocations occur each year in motor-vehicle crashes with a total annual cost 2.4 billion dollars. The objectives of this project are to definethe important biomechanical factors leading to these clinically observed disabling hip injuries, andto interpret and apply these hip-injury factors to define a comprehensive injury criteria for the entireknee-thigh-hip complex. This information will be used to improve the design of both anthropomorphictest devices (i.e., crash dummies) and vehicle interiors to help reduce the frequency of crash-inducedhip fractures.

Aortic ruptures have historically been a leading cause of death in motor vehicle accidents, secondonly to brain injuries. Aortic injuries are serious and often fatal with more than 80 percent of victimsexpiring before reaching the hospital. Despite its importance, the mechanisms of aortic injury are notwell understood and previous experimental efforts have not been successful in reproducing aorticinjuries in the laboratory. The UMTRI investigation will examine the relationships between organpositions and movements, lung pressure, and circulatory system pressure to aortic strains and failureof aortic tissue. Both static and dynamic testing is planned, with the goal of realizing an experimentalmodel that is capable of quantifying the biomechanical factors of aortic injuries.

Effects of Adjustable Pedals on the Distributions ofDriver Seat Positions and Eye Locations

In response to concerns about the proximity of drivers to steering wheel airbags, some vehicles arenow equipped with pedals that can be moved rearward on a motorized track to accommodate shorterdrivers. Ford Motor Company is sponsoring in-vehicle research to examine how drivers use the newadjustment feature. Do adjustable pedals produce the desired effect of increasing the spacing betweenthe driver and the steering wheel airbag? The principal investigator of the new project is Matt Reed,an assistant research scientist with UMTRI’s Biosciences Division.

The study will examine the pedal adjustment behavior and driving postures of people with a widerange of body dimensions. People who have experience with adjustable pedals in their own vehicleswill drive each of three Ford Expeditions equipped with different adjustable pedal configurations.Driving postures will be recorded with coordinate measurement equipment (FARO Arm) and thedrivers will rate several characteristics of the pedal systems.

Data analyses will determine the effects of pedal adjustment on the distributionsof driver-selected seat position and eye location (eyellipses). Regression methods developed in previ-ous UMTRI studies will be used to estimate percentiles of the seat position distributions for compari-son with the UMTRI seating accommodation model. Empirical eyellipses will be established

Reed

Van Ee


TECHNICAL REPORTS

Cole, D. E.; Londal, G. F. 2000. DELPHI X: forecast and analysis of the North American automotive industry. Volume 1:technology. Michigan University, Ann Arbor, Transportation Research Institute, Office for the Study of AutomotiveTransportation. 206 p. Report No. UMTRI-2000-3-1.

Eby, D. W.; Vivoda, J. M.; Fordyce, T. A. 2000. A study of Michigan safety belt use prior to implementation of standardenforcement. Michigan University, Ann Arbor, Transportation Research Institute, Social and Behavioral AnalysisDivision. 54 p. Sponsored by Michigan State Office of Highway Safety Planning, Lansing. Report No. UMTRI-2000-08.

Eby, D. W.; Fordyce, T. A.; Vivoda, J. M. 2000. Michigan safety belt use immediately following implementation of standardenforcement. Michigan University, Ann Arbor, Transportation Research Institute, Social and Behavioral AnalysisDivision. 63 p. Sponsored by Michigan State Office of Highway Safety Planning, Lansing. Report No. UMTRI-2000-25.

Joksch, H. C. 2000. Vehicle design versus aggressivity. Michigan University, Ann Arbor, Transportation Research Institute.136 p. Sponsored by Transportation Department, Research and Special Programs Administration, Cambridge,Mass. Report No. UMTRI-2000-22.

Kostyniuk, L. P.; Shope, J. T.; Molnar, L. J. 2000. Reduction and cessation of driving among older drivers in Michigan:final report. Michigan University, Ann Arbor, Transportation Research Institute, Social and Behavioral AnalysisDivision. 144 p. Sponsored by General Motors Corporation, Warren, Mich. Report No. UMTRI-2000-6.

Nowakowski, C.; Utsui, Y.; Green, P. 2000. Navigation system destination entry: the effects of driver workload and inputdevices, and implications for SAE recommended practice. Michigan University, Ann Arbor, TransportationResearch Institute, Human Factors Division. 69 p. Sponsored by Mitsubishi Electric Corporation, Amagasaki(Japan) Report No. UMTRI-2000-20.

Reed, M. P.; Lehto, M. M.; Flannagan, M. J. 2000. Field of view in passenger car mirrors. Michigan University, AnnArbor, Transportation Research Institute. 51 p. Sponsored by Michigan University, Ann Arbor, Industry AffiliationProgram for Human Factors in Transportation Safety. Report No. UMTRI-2000-23.

Rumar, K. 2000. Relative merits of the U.S. and ECE high-beam maximum intensities and of two- and four-headlampsystems. Michigan University, Ann Arbor, Transportation Research Institute, Human Factors Division. 51 p.Sponsored by Michigan University, Ann Arbor, Industry Affiliation Program for Human Factors in TransportationSafety. Report No. UMTRI-2000-41.

Sayer, J. R.; Mefford, M. L.; Huang, R. 2000. The effect of lead-vehicle size on driver following behavior. MichiganUniversity, Ann Arbor, Transportation Research Institute, Human Factors Division. 37 p. Sponsored by MichiganUniversity, Ann Arbor, Industry Affiliation Program for Human Factors in Transportation Safety. Report No.UMTRI-2000-15.

UMTRI RESEARCH NOTES

UMTRI BIBLIOGRAPHY

using analogous methods and compared both within vehicle (driver-selected vs. full-forward pedalpositions) and with respect to the UMTRI eyellipse model.

The results of the study will be used to improve the design of vehicles equipped with adjustablepedals. Models of the distribution of seat positions, eye locations, and clearances to the steering wheelwill provide an opportunity to improve interior component locations and restraint system design.


Schumann, J. 2000. Post-mounted delineators and perceptual cues for long-range guidance during night driving. MichiganUniversity, Ann Arbor, Transportation Research Institute, Human Factors Division. 21 p. Sponsored by MichiganUniversity, Ann Arbor, Industry Affiliation Program for Human Factors in Transportation Safety. Report No.UMTRI-2000-42.

Sivak, M.; Flannagan, M. J.; Schoettle, B.; Nakata, Y. 2000. Simultaneous masking of front turn signals by headlamps andother front lamps. Michigan University, Ann Arbor, Transportation Research Institute, Human Factors Division.149 p. Sponsored by Michigan University, Ann Arbor, Industry Affiliation Program for Human Factors inTransportation Safety. Report No. UMTRI-2000-44.

Streff, F.; Spradlin, H. 2000. Driver distraction, aggression, and fatigue: a synthesis of the literature and guidelines forMichigan planning. Michigan University, Ann Arbor, Transportation Research Institute, Social and BehavioralAnalysis Division. 40 p. Sponsored by Michigan State Office of Highway Safety Planning, Lansing. Report No.UMTRI-2000-10.

JOURNAL ARTICLES

Gillespie, T.; Karamihas, S. 2000. “Simplified models for truck dynamic response to road inputs.” Michigan University, AnnArbor, Transportation Research Institute, Engineering Research Division. 12 p. Heavy Vehicle Systems, Vol. 7, No.1, 2000, pp. 52–63.

Maio, R. F.; Shope, J. T.; Blow, F. C.; Copeland, L. A.; Gregor, M. A.; Brockmann, L. M.; Weber, J. E.; Metrou, M. E. 2000.“Adolescent injury in the emergency department: opportunity for alcohol interventions?” Michigan University, AnnArbor, Section of Emergency Medicine/ Michigan University, Ann Arbor, Transportation Research Institute, Socialand Behavioral Analysis Division/ Veterans Administration Hospital, Ann Arbor, Mich. 6 p. Annals of EmergencyMedicine, Vol. 35, No. 3, Mar 2000, pp. 252–257.

Nowakowski, C.; Green, P.; Kojima, M. 2000. “How to design a traffic-information web site: a human factors approach.”Michigan University, Ann Arbor, Transportation Research Institute, Human Factors Division. 11 p. ITS Quarterly,Vol. 8, No. 3, Summer 2000, pp. 41–51.

Sivak, M.; Flannagan, M. J.; Miyokawa, T. 1999. “Quantitative comparisons of factors influencing the performance of low-beam headlamps.” Michigan University, Ann Arbor, Transportation Research Institute, Human Factors Division. 9p. Lighting Research and Technology, Vol. 31, No. 4, 1999, pp. 145–153. Sponsored by Michigan University, AnnArbor, Industry Affiliation Program for Human Factors in Transportation Safety.

Waller, P. F.; Olk, M. L.; Shope, J. T. 2000. “Parental views of and experience with Michigan’s graduate licensing program.”Michigan University, Ann Arbor, Transportation Research Institute. 7 p. Journal of Safety Research, Vol. 31, No. 1,Spring 2000, pp. 9–15. Sponsored by National Highway Traffic Safety Administration, Washington, D.C.

CONFERENCE PAPERS

Fancher, P. S.; Baraket, Z. 2000. “Manual and automatic control of the longitudinal dynamics of individual motor vehicles.”Michigan University, Ann Arbor, Transportation Research Institute. 11 p. Froehling, L., ed. The Dynamics ofVehicles on Roads and on Tracks. Proceedings of the 16th IAVSD Symposium. Lisse, Swets and Zeitlinger, 2000.Pp. 270–281.

Winkler, C. B.; Ervin, R. D.; Hagan, M. R. 2000. “On-board estimation of the rollover threshold of tractor semitrailers.”Michigan University, Ann Arbor, Transportation Research Institute, Engineering Research Division. 11 p.Froehling, L., ed. The Dynamics of Vehicles on Roads and on Tracks. Proceedings of the 16th IAVSD Symposium.Lisse, Swets and Zeitlinger, 2000. Pp. 540–551.

UMTRI BIBLIOGRAPHY

To subscribe to the UMTRI Research ReviewComplete the form below and send it with a check for $35 made out to theUniversity of Michigan. This entitles you to a one-year subscription to theUMTRI Research Review.

Name _________________________________________ Date _________

Title _________________________________________________________

Organization __________________________________________________

Address ______________________________________________________

City ________________________________ State _____ Zip ___________

Mail your check for $35 and the form above to:

Monica Milla, EditorUMTRI RESEARCH REVIEWUniversity of Michigan Transportation Research Institute2901 Baxter RdAnn Arbor MI 48109-2150

Documents

Heavy Truck Stability