TECHNOLOGYREPORT published:17March2014 BIOENGINEERING AND BIOTECHNOLOGY doi:10.3389/fbioe.2014.00004 Investigation of theTHOR anthropomorphic test device for predicting occupant injuries during spacecraft launch aborts and landing JeffreyT.Somers1,NathanielNewby1,CharlesLawrence2,RichardDeWeese3,DavidMoorcroft3 and SheanPhelps4,5* 1ScienceTechnologyandEngineeringGroup,Wyle,Houston,TX,USA 2GlennResearchCenter,NationalAeronauticsandSpaceAdministration,Cleveland,OH,USA 3CivilAerospaceMedicalInstitute,FederalAviationAdministration,OklahomaCity,OK,USA 4GeorgiaTechResearchInstitute,Atlanta,GA,USA 5GeorgiaInstituteofTechnology,Atlanta,GA,USA Editedby: The objective of this study was to investigate new methods for predicting injury from TarunGoswami,WrightState expectedspaceflightdynamicloadsbyleveragingabroaderrangeofavailableinformation University,USA ininjurybiomechanics.Althoughallspacecraftdesignswereconsidered,theprimaryfocus Reviewedby: wastheNationalAeronauticsandSpaceAdministrationOrioncapsule,astheauthorshave HenriqueDeAmorimAlmeida, PolytechnicInstituteofLeiria,Portugal the most knowledge and experience related to this design.The team defined a list of AmyTeresaNeidhard-Doll,Wright criticalinjuriesandselectedtheTHORanthropomorphictestdeviceasthebasisfornew StateResearchInstitute,USA standardsandrequirements.Inaddition,theteamdown-selectedthelistofavailableinjury *Correspondence: metrics to the following: head injury criteria 15, kinematic brain rotational injury criteria, SheanPhelps,HealthSystems, GeorgiaTechResearchInstitute,260 neckaxialtensionandcompressionforce,maximumchestdeflection,lateralshoulderforce 14thStreetNW,Suite455,Atlanta, and displacement, acetabular lateral force, thoracic spine axial compression force, ankle GA30332,USA moments,andaveragedistalforearmspeedlimits.Theteamfeltthatthesemetricscapture e-mail:[email protected] alloftheinjuriesthatmightbeexpectedbyaseatedcrewmemberduringvehicleaborts and landings. Using previously determined injury risk levels for nominal and off-nominal landings, appropriate injury assessment reference values (IARVs) were defined for each metric.Musculoskeletaldeconditioningduetoexposuretoreducedgravityovertimecan affectinjuryriskduringlanding;thereforeadeconditioningfactorwasappliedtoallIARVs. Althoughthereareappropriateinjurydataforeachanatomicalregionofinterest,additional researchisneededforseveralmetricstoimprovetheconfidencescore. Keywords:injurycriteria,spaceflight,dynamicloads,anthropometrictestdevice,testdeviceforhumanoccupant restraint,injuryassessmentreferencevalues INTRODUCTION develop a standardized test methodology (i.e., ATD, seat, suit, PURPOSE acceleration profiles, etc.) for inclusion in National Aeronautics The objective of this work was to: (1) identify a list of criti- andSpaceAdministration(NASA)Standard3001,andallrelated cal spaceflight injuries from dynamic loading that need to be programrequirements(NationalAeronauticsandSpaceAdmin- protectedagainsttoenablemissionsuccess,(2)identifyananthro- istration,2011). The information included is based on available pomorphic test device (ATD) to be used to predict the thresh- data at the time of the report and upon opinions from experts old at which human injuries will occur, and (3) develop a at NASA, the National Highway Traffic Safety Administration table of ATD thresholds known as injury assessment reference (NHTSA),and the FederalAviationAdministration (FAA). The values (IARVs) for each critical injury. The eventual goal is to teamconsideredoccupantprotection(OP)needsduringdynamic Abbreviations:AIS,abbreviatedinjuryscale;ATD,anthropomorphictestdevice; BDRC, Brinkley dynamic response criterion; BrIC, brain rotational injury cri- Center;KSC,KennedySpaceCenter;LEO,lowearthorbit;LM,LockheedMartin; teria; CAMI, Civil Aeromedical Research Institute; CG, center of gravity; CM, LOC,lossofconsciousness;MPCV,multi-purposecrewvehicle;mTBI,mildtrau- command (crew) module; CSDM, cumulative strain damage measure; DAR, maticbraininjury;NASA,NationalAeronauticsandSpaceAdministration;NBDL, definition of acceptable risk; EAFB, Edwards Air Force Base; EVA, extravehic- Naval Biodynamics Laboratory; NESC, NASA Engineering and Safety Center; ular activity; FAA, Federal Aviation Administration; FE, finite element; FEM, NHTSA,NationalHighwayTrafficSafetyAdministration;OP,occupantprotec- finite element modeling; FTSS, First Technology Safety Systems; GRC, Glenn tion;PMHS,post-mortemhumansurrogates;SID,side-impactdummy;SLS,space Research Center; HIC, head injury criteria; HRP, Human Research Program; launchsystem;SRB,solidrocketboosters;TBI,traumaticbraininjury;TBD,to HSIR, human-systems integration requirements; IARVs, injury assessment bedetermined;THOR,testdeviceforhumanoccupantrestraintmodificationkit; reference values; IRL, Indy Racing League (aka: INDYCAR™); ISS, interna- USAARL,United StatesArmyAeromedical Research Laboratory; USSR,United tionalspacestation;JARI,JapanAutomobileResearchInstitute;JSC,JohnsonSpace SovietSocialistRepublic;WSTF,WhiteSandsTestFacility. www.frontiersin.org March2014|Volume2|Article4|1 Somersetal. THORATDIARVsforspaceflight phasesofspaceflight,whichincludeabort(padabortandascent abort)aswellasre-entryandlanding.Althoughvariousspacecraft designswereconsidered,theprimaryfocuswastheNASAOrion capsule,astheauthorshavethemostknowledgeandexperience relatedtothisdesign. SPACEFLIGHTDESIGNCONSIDERATIONS Vehicledesigns Reachingspacerequiresanextremeamountofkineticenergy,and effective systems to dissipate this energy on the return to Earth. Whilemostofthisenergyiscontrolled,dissipated,orabsorbedby thevehicle,someamountofkineticenergymaybetransmittedto theoccupantsaboardthespacecraft.Thisenergy,ifnotproperly managed,maycauseinjurytothecrewmembers.Vehicledesignis animportantconsiderationformanagingthisenergy,particularly duringlaunchabortsandlanding. Launchandabortsystems. Currently,chemicalrocketsareused to launch humans into space. These systems typically accelerate the crew vehicle to orbital velocities >6,900m/s (15,430mph) FIGURE1|Nominalcapsulelandingorientationwithrespecttowater to attain low earth orbit (LEO) within 10min of launch. These (landlandingsimilar). sustainedaccelerationsaredesignedtobewellwithinhumantol- erance. Because of the amount of energy stored in the launch vehicle(eitherliquidorsolidpropellant),therearefailuremodes positionedreclined90°fromverticalontheirback,contactingthe thatnecessitatethedesignofabortsystems. impact surface feet first, resulting in a combined +X (eyeballs Mosthumanspaceflightvehiclesdesignedtodatehaveincluded in)and+Z (eyeballsdown)primarylandingload(seeFigure1), launchphaseabortcapabilities.FortheMercuryandApollopro- althoughlandingdynamicsareheavilydependentonthespecific grams in the U.S. – as well as the Soyuz program in Russia – a designandfailuremodes. launch escape system was included in the spacecraft design to TheMercury,Apollo,multi-purposecrewedvehicle(MPCV), allow quick separation of the crew module away from the main andSpaceXDragonaredesignedtolandprimarilyintheocean, vehicleincaseofacatastrophicfailureofthelaunchsystem. but can also land on land in contingency cases. The Vostok, For the Russian Vostok and Buran programs, as well as the Voskhod, Soyuz, and Boeing CCT-100 are all designed for pri- U.S.Geminiprogram,ejectionseatswereincludedinthespace- mary landing on land, with contingency landing capability in craftdesigntoallowcrewmemberstoescapeseparatelyfromthe water.Capsulesmayincludeadditionalenergymanagementfea- entirelaunchvehicle,althoughtheycouldonlybeoperateddur- tures such as retrorockets, airbags, energy-absorbing structures, ingaveryshortperiodofthelaunchprofile.Noabortcapabilities and/orstrokingseats. existedoutsideofthisperioduntilsufficientaltitudewasreached toallowforanormalseparationanddescent.TheU.S.SpaceTrans- Liftingbodydesign. Aliftingbodydesignimprovesuponthecap- portation System Program (the formal name for NASA’s“Space suledesignbyaddingliftingsurfacestoincreasemaneuverability Shuttle” program) included four primary elements: an orbiter andcross-rangeperformance.SierraNevada’sDreamChaserisan spacecraft (Space Shuttle), two solid rocket boosters (SRB), an exampleof aliftingbodydesign,andisbasedonNASA’sHL-20 externaltankhousingfuelandoxidizer,andthethreeSpaceShut- prototype(NationalAeronauticsandSpaceAdministration,2013; tle main engines. The Space Shuttle included ejection seats that SierraNevadaCorporation,2013).FromanOPstandpoint,alift- were disabled after the first four flights and eventually removed ingbodyrepresentsalessviolentlandingenvironment,asitdoes (Jenkins,1999). notrelysolelyonparachutestodissipatelandingvelocity.Evenin All future NASA vehicles are required to have a crew escape off-nominallandingconditions,thisdesignfeatureisexpectedto system(NationalAeronauticsandSpaceAdministration,2012a). reducethepotentialforinjury. Becausecrewescapesystemsmustquicklyseparatethecrewaway fromthelaunchvehicle,thecrewmaybeexposedtohighdynamic Landingloadsandvectors loads.Theseloadswillvarydependingonvehicledesign,phaseof Becauserelativelylittleisknownaboutlandingloadmagnitudes flight,andothervehicleperformancecharacteristics. and direction vectors for future commercial vehicles (SpaceX Dragon, Boeing CCT-100, and Sierra Nevada Dream Chaser), Capsule landing systems. The capsule design was the original knowledge of the MPCV design was used as a logical basis for designchosenforhumanspaceflightprimarilybecauseitismass this work. This is likely the bounding case for all of the com- efficient and simple. To date, this design has been used for all mercialvehiclesforseveralreasons.First,theMPCVislargerand humanspaceflightprograms,withtheexceptionoftheU.S.Space heavierthanitscommercialcounterparts,whichshouldresultin Shuttle.Capsulevehiclestypicallylandwiththecrewinaseated larger landing loads than the commercial vehicles. Second, the FrontiersinBioengineeringandBiotechnology|Biomechanics March2014|Volume2|Article4|2 Somersetal. THORATDIARVsforspaceflight MPCV was evaluated for its performance during land landings. Although Orion used a slightly different method of defin- Although the landings were severe, they were not significantly ing nominal and off-nominal,based on the probabilities of off- worse than some off-nominal water landing scenarios. In addi- nominallandingsandtheassociatedlevelsofriskforeach,thetotal tion,the MPCV was not designed for a primarily land landing, landingcasesthatare≤µ+1.5σthresholdareanapproximation sovehiclesthataredesignedprimarilyforlandlanding(likethe oftheOrionrisklevelfornominal,andthelandingcasesbetween CCT-100) would likely have a softer land landing. Finally, for µ+1.5σandµ+2.5σaresimilartotherisklevelforoff-nominal. the lifting body designs, the MPCV water impact will be more Allcasesaboveµ+2.5σlandingcasewouldbecontingencyand severethanoff-nominallandingsencounteredduringahorizontal not considered for design purposes (0.6% of cases). Using this landing. definition,an off-nominal event will occur approximately every For the MPCV,the landing loads are primarily +X (eyeballs 4yearsandacontingencyeventevery40years,oreffectivelynever in)and+Z (eyeballsdown).Thisisasimilarorientationtothe (assumingfourlaunchesperyear). ApolloCM(seeFigure1).The+X componentisprimarilydue toverticalvelocityof thevehicleandisdependentonparachute Crewdeconditioning performance.Inthecaseofatwo-parachutelanding,thevertical Severalphysiologicchangesoccurinresponsetomicrogravity.The velocity will be higher, resulting in a higher +X load. The +Z twomainchangesconcerningimpacttolerancearebonemineral componentisdrivenbytwofactors.First,thehorizontalwindand densitylossandmusculartissueatrophy. wave speed affects the horizontal velocity. At high wind speeds Duringprolongedspaceflight,skeletaldensitychangesprimar- the+Z loadismuchhigher.Second,thevehicleisdesignedwith ilyinthelowerextremitiesandspine(Langetal.,2004)consistent a set hang angle,tilting the entire capsule so that only the edge with Wolff’s law (Wolff, 1986). Studies conducted using dual ofvehiclecontactsthewaterfirst.Evenwithnohorizontalwind, energy X-ray absorptiometry (DXA) have shown bone mineral thishanganglewillimparta+Z axisloadbasedonthevertical densitydecreasesonaverageof1–1.6%inthespine,femoralneck, velocity. In addition,wave slope during water landings and ter- trochanter,and pelvis,with an average loss of 1.7% in the tibia rainslopeduringlandlandingscontributetotheeffectiveimpact after only 1month in microgravity (LeBlanc et al., 1998; Vico angle. et al.,2000). Because skeletal deconditioning is time dependent, Althoughthesevectorsareprimarytothelandingload,theload anymethodforaccommodatingthelosseswillbemissionlength isalsodependentonthevehiclemaintainingthecorrectorienta- specific. tion,so that the crew are oriented correctly at landing (i.e.,feet Changesinmusclemassandstrengthalsooccur,andaredepen- first).Anonboardcontrolsystemisneededtomaintainthisori- dentontheexerciseregimeemployedduringspaceflight.During entation.Withoutacontrolsystem,thevehiclecanrotatebefore Skylabmissions,legvolumedecreasedby7–10%(Thorntonand impact, imparting loads in other directions. The landing loads Rummel,1977)andupto19%increwmembersaboardtheMir may be quite complex,necessitating consideration of all vehicle spacestation(Stein,1999;LeBlancetal.,2000).Themuscleloss dynamics. experienced by crewmembers is also selective; muscle fiber size in the vastus lateralis decreased after 5–11days in flight at dif- Landingmodes ferentrates.Edgertonetal.(1995)andZhouetal.(1995)found Each spaceflight vehicle design has a unique launch, abort, re- decreases of 16% in Type I,23% in Type IIa,and 36% in Type entry, and landing environment. Each vehicle is optimized for IIbfibers.Intermsofmusclestrengthduring6monthmissions, itsparticular“nominal”caselanding,orthelandingthathasthe astronautsexperiencedmusclestrengthlossupto24%attheknee highestprobabilityofoccurring.Todeterminethislandingmode, andupto22%intheankle(Gopalakrishnanetal.,2010). detailed analyses of the vehicle systems and environmental fac- Itiscurrentlyunclearhowthesechangesaffecthumanphysiol- tors are conducted. These analyses identify the distribution of ogyandimpacttoleranceinthesettingofspaceflightandlanding all possible landings related to the normal and tangential veloc- following long duration mission profiles. However, it is fairly ities. Assuming a normal distribution of all landing probabili- evident that greater OP measures will be needed for decondi- ties,thresholdsfornominalandoff-nominalcanbedefined(see tioned crewmembers. Currently, these effects are accounted for Table1).Itshouldbenotedthatforcapsule-basedvehicles,even byapplyinglowerdynamicloadlimits,whicharebasedonNASA’s nominallandingdynamicsaremorelikeanautomobileaccident, IntegratedMedicalModel(Lewandowskietal.,2008).TheInte- thannormalautomotiveaccelerations. gratedMedicalModelofbonelossisderivedfrombonemineral Table1|Comparisonofpossiblenominalandoff-nominaldistributionthresholds. Threshold Percentageof Percentageofcases Exp.freq. Approx.occurrenceforonelandingoutside Approx.designlevels landingcases outsideofrange outsideofrange ofrange(assumingfourflightsperyear) forOrion ≤µ+1σ 84.1 15.9 1in6 Every18months ≤µ+1.5σ 93.3 6.7 1in15 Every4years Nominal ≤µ+2σ 97.7 2.3 1in44 Every11years ≤µ+2.5σ 99.4 0.6 1in161 Every40years Off-nominal ≤µ+3σ 99.9 0.14 1in769 Every200years www.frontiersin.org March2014|Volume2|Article4|3 Somersetal. THORATDIARVsforspaceflight density changes, not on actual measured and scientifically vali- this phase of flight. Additionally, the launch, launch abort, and datedhumanimpacttoleranceintheseconditions.Thisapproach landingenvironmentsforNASAareextremeinnaturecompared maybeacceptableforshortstaysontheinternationalspacesta- withwhatisnominallyexperiencedduringautomobiledrivingor tion(ISS),butmaynotapplytoorprotectagainstthedeleterious civilian/militaryflight.Thedifferencesbetweenautomotive,mil- physiological effects of longer-duration missions to near-Earth itary,andNASAoperationalenvironmentsmakeitimportantto objects,themoon,and/orMars,sincethereislittleknownabout considertheoverallacceptableprobabilityofinjury,giventhatoccu- spaceflight deconditioning beyond 6months to 1year. In addi- pantsinaspacevehiclewillincurimpactconditionseverytime tion,inflightcountermeasuresarebeingdevelopedtocounteract thevehiclereturnstoEarth. thesephysiologicalchanges,sointhefuture,loweringtheresponse Suitconsiderations limitsmaynotbenecessary. OneoftheuniqueaspectsoftheNASAenvironmentisthepres- sureorspacesuit.Thissuitprotectsthecrewfromthevacuumof Injuryriskposture spacebyprovidinga pressurizedenvironmentaroundthebody, TogaininsightintowhatNASA’sinjuryriskpostureshouldbe,it abreathableatmosphere,thermalprotection,andmicrometeorite ishelpfultoreviewotherindustriesandtheirrespectiveriskpos- protection(whenoutsidethevehicle).Inadditiontothesebasic tures derived from their contextual operational scenarios based functions,otherconsiderationsinsuitdesignincludemobility,fit uponscientificevidencegatheredtherein. onawiderangeofcrewmembers,andcontingencyextravehicular Fortheautomotiveindustry,specificallypassengercars,most activity (EVA). With all of these demands on the suit, the final injurylimitsarebasedona5–50%riskof anabbreviatedinjury scale(AIS)3+injury,whichdelineatestheoccurrenceofasevere design is often not optimized for OP. There are several consid- erations for the occupant during abort and landings that relate injury(AssociationfortheAdvancementofAutomotiveMedicine, to suit design. First,the suit,unlike most clothing,may contain 2005).Althoughthisseemslikeanobjectionableriskposture,there rigid elements. Depending on the placement of these elements, aretwomainreasonsthatthispostureisacceptablefortheauto- point-loads or blunt trauma may occur resulting in crew injury motiveindustry.First,theselimitsarebasedonstandardizedtests (McFarlandandDub,2010).Therefore,theplacementanddesign that represent a worst case scenario, and not a“representative” ofthesecomponentsarecriticaltoprotectingthecrew(Danelson collision. Second, the overall probability of a person being in a etal.,2011).Second,head-mountedmasscanposeaseriousthreat crashanytimeheorshegetsintoavehicleisveryremote(1in tocrewmembersiftheadditionalmassiscarriedbytheneck(Rad- 120,000) for passenger vehicle usage (National Highway Traffic fordetal.,2011).Finally,becausethesuitisapressuregarment, SafetyAdministration,2007,2009). Therefore,after considering thereisachanceoflandingwiththesuitinflated.Inthiscase,the theseverityofthecollisionandtheprobabilityofactuallyhaving vehiclerestraintsystemisnolongerrestrainingthecrewmember, acollision,thetotalriskofinjurytoautomobiledriversisverylow. butisinsteadrestrainingthesuit.Insidethesuit,thecrewmember For military aircraft, the situation is similar, although more maybefreetomove,increasingthepossibilityofinjury.Inall,the inherentriskisinvolved.Militaryaircraft(bothfixedandrotary suitpresentsauniquechallengethatmustbeaddressedtoprevent wing)aredesignedtoallowforahigherriskposturethanwhatis crewinjury. desiredandacceptableforspacecraftoccupantsassignedtoNASA. Again,thesehigherlevelsof riskareconsideredacceptablegiven Genderandanthropometrics thattheoverallriskofinjuryper“sortie”(definedasonedeploy- UnlikepreviousNASAcapsuledesigns,futureNASAvehiclesmust mentordispatch–launchwithsubsequentlanding–ofamilitary becapableof accommodatingmenandwomeninawiderange aircraftonamission)is1in670(orless),eventhoughthisissig- ofanthropometrics.CurrentNASAvehiclerequirementsstipulate nificantlyhigherthantheriskseenintheoperationofpassenger a vehicle must accommodate a 1st percentile female to a 99th cars(Mapes,2006). percentile male. Protecting for such a wide range of sizes is a ForNASA,thesituationisverydifferent.Withpassengervehi- challenge. Most OP data are based either on young,male,mili- cles,millionsofmilesaredriveneachyearwitharelativelylowrisk tarysubjects,oronelderlymalepost-mortemhumansurrogates ofcollisionorinjury.Forthemostpart,passengervehicleriskis (PMHS).Asof2011,theastronautcorpsmedianageis47.1years constantduringtheentiretripor“sortie.”Similarly,formilitary for males (range 35–56) and 43.3years for females (range 32– aircraft,thousands of flight hours are logged with relatively low 52), with males accounting for 76% of the corps. In terms of risksofinjury,andlikepassengervehicles,thereissignificantrisk anthropometry,statureis177.3±4.9cm(4th–95thpercentile)for during the entire mission although this risk is a direct result of male crewmembers, and 168.9±4.3cm (25th–97th percentile) differentvariablecauses(enemyfire,mechanicalfailures,weather, for female crewmembers. For weight, male crewmembers are piloterrors,humanfactors,etc.).However,unlikepassengervehi- 79.3±6.9kg (6th–75th percentile), and 63.2±8.9kg (4th–65th cles,militaryaircrewsaresubjecttoahigherriskofinjuryduring percentile) (as shown in Table2). Because of the wide range of takeoff and landing. This risk is closer to NASA’s environment demographics, accurately determining injury risk for the entire where risk of injury due to dynamic loads is concentrated dur- rangewithintheastronautcorpsisdifficult,andtheresultswill inglaunchandlandingphasesofoperationswhenflightproduces containacertainamountofuncertainty. the highest loads on the vehicle. Unlike passenger vehicles and military aircraft, there is very low risk of injury due to impact CURRENTNASASTANDARDS onceastableorbithasbeenreached,dueprimarilytothefactthat Currently, NASA standards for transient accelerations (≤0.5s) thereareverysmallloadsappliedtothevehiclestructureduring are based primarily on the Brinkley dynamic response criterion FrontiersinBioengineeringandBiotechnology|Biomechanics March2014|Volume2|Article4|4 Somersetal. THORATDIARVsforspaceflight Table2|NASAastronautcorps,gender,andanthropometric • Phase 5 – injury risk functions and NASA standards develop- distribution(asof2011). ment. Gender Medianage Percentage Stature Weight OnlyPhase1oftheforwardplanisdiscussedinthisreport. (range) ofcorps (percentile) (percentile) LITERATUREREVIEWANDSTANDARDSFRAMEWORK Male 47.1(35–56) 76 177.3±4.9cm 79.3±6.9kg DEVELOPMENT (4th–95th) (6th–75th) Basedontheresultsof theexpertpanelsummit,workbeganto Female 43.3(32–52) 24 168.9±4.3cm 63.2±8.9kg determinealistof criticalinjuriesthatneedtobemitigated,the (25th–97th) (4th–95th) bestATDforpredictingbiodynamicresponses,andIARVneeded tomitigatetheinjuryrisk.Basedoncurrentliterature,aframe- work was developed for completing these tasks and identifying (BDRC)(NationalAeronauticsandSpaceAdministration,2011). areasinneedofadditionalresearch. The BDRC is a simple,lumped-parameter,and single degree of freedom model that estimates the whole body response due to CRITICALINJURYDEFINITION appliedacceleration,andiscomputationallyefficientandrequires Theexpertpanelidentifiedtheneedforaconciselistof“critical” verylittleintermsofvalidationtesting. injuriesthatNASAwouldneedtomitigate.Entriescouldbebased However, the BDRC has limitations (Somers et al., 2013). It onlikelihood(basedonpreviousspaceflightexperience)orthose onlypredictsrangesof injuryriskandcannotprovideinforma- thatcouldcausethecrewtobeincapableofperformingpostflight tion as to the severity or anatomical location of an injury. Only tasks. The list is not all-inclusive,but is intended to protect the the+Z axisinjuryriskiswellvalidated(withoperationalejection crew from injuries that if successful,would also allow a level of seatdata);the±X,±Y,and−Z injuryrisklevelsarestatistically protectionfrommoresevereinjuries.Forexample,ifNASAwere limited (Brinkley,1985). In addition,these data do not account toprotectagainstribfracturesandlungcontusions,theassump- for the interactions between crewmembers, suit, and seat. The tionisthatotherinternalorganswouldbeprotectedtothesame seat and helmet used in the human testing and development of levelaswell. the BDRC are very different from what is actually planned for ThelistofinjurieswasusedtoformthebasisoftheATDselec- MPCV. Previously performed tests did not typically include a tion.TheATDneededtobecapableofgeneratingaresponsethat suit,sosuitinteractionsarenotaccountedforintheinjuryrisk isrelevanttothetypesof injuriesNASAwishestomitigate.The prediction. injurylistcombinedwiththeselectedATDwouldultimatelydrive Given the limitations of the current NASA Standards, the theselectionoftheinjurymetricsandIARV. humanresearchprogram(HRP)andtheNASAEngineeringand Before establishing a critical injury list, several assumptions SafetyCenter(NESC)beganstudyingalternativesforinclusionin weremadebasedontheexistingNASAstandardsasfollows: thestandard.Thisstandardupdateworkisprimarilyfocusedon theOrionvehicle. • Afive-point(orbetter)racingharnesswouldbeused. • Transient accelerations (<0.5s) would not exceed a moderate EXPERTPANELSUMMIT injuryrisklevels. AfterworkingwiththeOriondesignexclusively,theOPteamheld • Sustained (>5s) linear accelerations, rotational accelerations, anexpertsummitinHoustoninJune2010.Thegoalofthesum- rotationalvelocities,andaccelerationrateofchangewouldnot mit was to develop an OP plan that would not only further the exceed levels specified in the standard (NASA Standard 3001) Orioneffort,butalsoadapttothecommercialcrewvehicles. (NationalAeronauticsandSpaceAdministration,2011). Experts from the U.S. Army, U.S. Navy, U.S. Air Force, • Transientrotationalaccelerationswouldnotexceedlevelsspeci- FAA, NHTSA, Indy racing league (IRL) (INDYCAR™), univer- fiedinthestandard(NASAStandard3001)(NationalAeronau- sity researchers fromVirginia Tech,Wake Forest University and ticsandSpaceAdministration,2011). Wayne State University, and automotive biomechanics experts • Minimalornobodymovement(basedontheBrinkleyAmpli- attendedthesummit.NASApersonnelfromseveralareasinclud- ficationRule). ing the Human Health and Performance Directorate (HRP, the • Thevehiclewouldmaintainanoccupantsurvivablevolume. Biomedical Research and Environmental Sciences Division,and • Requirements would be met as for any other vehicle such as theSpaceandClinicalOperationsDivision),theAstronautOffice, sharpedges,pinchpoints,etc. theSafetyandMissionAssuranceDirectorate,andtheNESCalso participated. Some assumptions were developed that were thought to be Groupconsensuswasreachedontheforwardplanconsisting generalenoughtoencompassmostfutureNASAvehicles,yetcon- ofthefollowingelements: strained enough to allow a useful set of injuries to be defined. Theseare: • Phase1–literaturereviewandstandardsframework. • Phase 2 – ATD testing and finite element model (FEM) • Crewmembers may be in a pressurized suit,an unpressurized assessment. suit,orunsuited. • Phase3–humanexposuredatamining. • Thecrewwouldberecoveredwithin24h(thisisbasedonanaly- • Phase4–humantestingandcorrelationtoATDresponse. ses conducted during NASA’s Constellation Program showing www.frontiersin.org March2014|Volume2|Article4|5 Somersetal. THORATDIARVsforspaceflight that most locations worldwide can be reached within this interest and concern for returning crewmembers, as spaceflight timeframe). deconditioningcanaffecthipstrength.Hiploadswereaddedto • Thecrewmustbeprotectedtotheextentthatallpost-landing the matrix (see Table3) after the initial discussion with experts egresstaskscanbecompleted. basedontheriskoffemoralheadfractureduetolateralloads. • Crewmembertasksrequiredtoegressthevehiclepost-landing areassumedtobesimilarinhumanperformancetoOrion(i.e., INJURYMETRICANDANTHROPOMORPHICTESTDEVICESELECTION similar physical abilities needed, but not necessarily the same Following the expert panel recommendations,a team was orga- tasks). nized to determine the best ATD (and associated metrics) for • Dynamicloadsexperiencedbythecrewwouldbelessthanor validating against the critical injury list. The team (consisting equaltothecurrentpredictionsfortheOrionvehicle(withinthe of researchersfromNHTSA,FAA,andNASA)beganevaluating sameorderofmagnitudeofthecurrentOrionassumedloads). whichATDmetricswerethebestchoicesgiventhelistofcritical • Design of the vehicle would prevent inadvertent contact with injuries and NASA’s dynamic environment. The team examined vehicleinteriorexcludingtheseatandsuit(i.e.,seatcanstroke eachanatomicalregionandidentifiedavailableinjurymetricsthat into other structures, knees can’t contact control panel, and couldaddresseachcriticalinjury.Eachmetricidentifiedwasasso- stoweditemswillnotbefree). ciatedwithanATDthatprovidesthatmeasurementcapability(see • Only considering injuries induced by dynamic loads (i.e.,not Table3). consideringinhalationdangers,fire,etc.). BecauseIARVsarerelatedtotheparticularATDemployed,the teamconsideredalloftheapplicableATDsavailableincludingthe Based on these assumptions, the team identified the critical Hybrid-III, the WorldSID, and the test device for human occu- injuries in several regions including the head,face,chest,upper pantrestraint(THOR).Duringdiscussionsoftheinjurycriteria, extremities,lowerextremities,andthespine.Forthehead,concus- the team reviewed the benefits and drawbacks of each ATD as sionwithandwithoutthelossofconsciousness,skullfracture,and shown in Table 4. The final consensus was to use the 50th per- traumaticbraininjurywereallidentifiedbythepanelascritical centileTHORATDastheprimaryATDfortestingandanalysis. injuries.Fortheface,eyeandearinjuries,andfacialfractureswere Although it is only currently available in one size and the cur- allidentified.Lungcontusions,ribfractures,hemo-,pneumo-,and rentFEMisstillpreliminary,theteamchosetheTHORbecause hemopneumothoraxwereclassifiedascriticalchestinjuries.For of its biofidelity and multi-axis performance. To overcome the upperandlowerextremities,jointinjury(includingshoulderdis- limitations of the THOR, IARVs will eventually be chosen that location)andskeletalfracturewerecategorizedascriticalaswell. incorporate the increased risk of gender differences and various Finally, for the spine (cervical, thoracic, and lumbar), brachial anthropometries.Inaddition,lateraltestscomparingthe50thper- plexusinjury,cordcontusion,vertebralfracture,herniateddisks, centileWorldSID and the 50th percentile THOR are planned to anddiskrupturewerealladdedtothelist. determineif theTHORisapragmaticchoiceforlateraltesting. After working with the expert panel, and in the process of AlthoughtheTHORwasnotdesignedforside-impactconditions, testing, the team determined that hip loads are of particular the IRL has conducted side-impact testing using the THOR. If Table3|Matrixshowingtheinjurymetricchosenandtherelationshipwiththecriticalinjuries. y y ur ur Headinjury Facialtrauma Cervicalspinetrauma Blunttrauma Lungcontusion Ribfracture Hemopneumothorax Upperextremityjointinj Upperextremityfracture Femoralheadfracture Thoracicspinetrauma Lumberspinetrauma Lowerextremityjointinj Lowerextremityfracture HIC15 T/H BrIC T/H Neckaxialtension T/H Neckaxialcompression T/H Maxchestdeflection T T T Lateralshoulderforce(deflection) T/W T/W T/W T/W T/W Acetabularlateralload T/W Thoracicspineaxialcompression T/H T/H Anklemoments T Contactlimits/restrains(designconstraint) X X X X X X H,Hybrid-III;T,THOR;W,WorldSID;X,designconstraint. FrontiersinBioengineeringandBiotechnology|Biomechanics March2014|Volume2|Article4|6 Somersetal. THORATDIARVsforspaceflight Table4|BenefitsandlimitationsofeachATD. Hybrid-III THOR WorldSID Benefit Limitation Benefit Limitation Benefit Limitation Injurycriteria Injurycriteria Notrepresentativeof Improvedneck Severalinjurycriteria Improvedshoulder Severalinjury readily humanresponses biofidelity stillindevelopment biofidelityover criteriastillin available Simulatedneckmuscle EuroSID development tension Injurycriterianot Improvedchest Improvedchest developedforlow geometryandresponse biofidelityover injuryrisk Betterfrequency EuroSID response Betterresponseatlow dynamics FEmodel FEmodel Notwellvalidatedin FEmodelispreliminary NoFEmodel status commercially otherloading anddoesnotreflect available available directions currentdesign Anthropometry 5th,50th, 5tha,50th Onlyone 5tha,50th Onlyone 95th anthropometricsize anthropometric currentlyavailable sizeavailable Directionality Validatedin Requiresmodification Validatedinfrontal Unknownresponsesin Validatedin NotsuitedforX frontal forZ axisb impacts Y andZ axesc side-impacts axis impacts NotsuitedforY axis ClosestATDto multidirectional Availability Readily Limitedavailability Limited available availability aFifthpercentileTHORandWorldSIDcurrentlyindevelopment.WillnotbeavailableforNASAuseforseveralyears. bZaxistestingwiththeHybrid-IIIrequiresmodification(“Aerospace”model). cUsedbytheIndyracingleague(IRL)inside-impacttests.Furthertestingrequired. possible,the team would prefer to use the THOR in all axes to Table5|Injuryclassificationmappingandacceptablerisklevels. simplifyanalysisandtesting;however,furtherevaluationsof the THOR and WorldSID in lateral loading conditions will be nec- Injuryclass Acceptablerisklevel essary to validate this approach. In the event that the THOR acrossallmissions is found to be inadequate in predicting injuries due to lateral Nominal(%) Off-nominal(%) loads,theteamwouldrecommendusingtheWorldSIDforsuch cases. ClassI(AIS1+) 5 19 ClassII(AIS2+) 1 4 INJURYASSESSMENTREFERENCEVALUESASDETERMINEDFROMA ClassIII(AIS3+) 0.3 1 LITERATUREREVIEW ClassIV(AIS4+) 0.03 0.1 Having selected the ATD and injury metrics, work focused on the IARV associated with each metric. Since most available lit- eratureclassifiesinjuriesaccordingtotheAIS,theteamdecided IARVforheadinjurycriteria tomaptheAISleveldirectlytoeachinjuryclassification(Associ- Theheadinjurycriteria(HIC)arethestandardheadinjurypre- ationfortheAdvancementof AutomotiveMedicine,2005).The dictorintheautomotiveindustry.Equation1isusedtocalculate teamusednominalandoff-nominalrisklevelsfromthedefinition theHIC(PrasadandMertz,1985). ofacceptablerisk(DAR)todeterminetheappropriateIARVsfor Headinjurycriteriaformula: eachmetricasshowninTable5(Somersetal.,2014).Inaddition, fitoartievaechesItAimRaVt,eaocfotnhfiedceonncfiedsecnocreeiinsitnhceluIAdeRdV.sTrheipsosrctoerdeaisnadqisuoaln- HIC = max (cid:32)(t −t )(cid:20)(cid:90) t2a(t)dt 1 (cid:21)2.5(cid:33) (1) ascaleof0–5(noconfidence–fullconfidence,respectively). 15 0≤t2−t1≤0.015 2 1 t1 t2−t1 www.frontiersin.org March2014|Volume2|Article4|7 Somersetal. THORATDIARVsforspaceflight Recentstudiesofmildtraumaticbraininjury(mTBI)infoot- ATDused,soisapplicabletotheTHOR.TheBrICiscalculated ball players can be very useful for determining the appropriate usingEq.4. thresholdforheadinjury.SinceAIS1and2injuriestothebrain BrICformula: areofprimaryconcern,theHICinjuryriskfunctionsfromFunk (cid:115) Vetiragl.in(2ia0T0e7c)hwdilaltbaereupsoedrte(Edqb.y2F).uTnhkeisne2d0a1ta2wbeecraeucsheotsheenHoIvCervtahle- BrIC= (cid:18)ωωx (cid:19)2+(cid:18)ωωy (cid:19)2+(cid:18)ωωz (cid:19)2 (4) xC yC zC uesfromthe2007studyaremoreconservative(Funketal.,2012). BecauseconcussioninjuryriskdeterminedfromNASCARhead where,ω is the maximum angular head velocity in the i plane, injury modeling resulted in much higher allowable HIC values ω istheiangularheadvelocitycriticalvaluefortheiplane. (Somersetal.,2011),theFunkHIC15curvewillbeusedtobe iCThe critical values are 66.3,53.8,and 41.5rad/s for the X,Y, conservativeuntiltheNASCARresultscanbeverifiedwithaddi- andZ axes.Equation6detailshowtodetermineIARVsforBrIC. tional datasets. In addition,the THORATD is assumed to have Figure3showstheresultantinjuryriskcurves. similar head kinematics as the Hybrid-III ATD. That allows for BrICinjuryriskmodel: theuseofidenticalHIClimitsforbothATDs.Figure2showsthe resultantHIC15injuryriskfunction. (cid:16) (cid:17)2.84 Headinjurycriteriainjuryriskmodel: p(AIS≥n|BrIC)=1−e− BλrInC (5) (cid:16) (cid:17)α p(cid:0)inj|AIS≥1(cid:1)=1−e− HβIC (2) BrICIARVdetermination: (cid:113) HeadinjurycriteriaIARVmodel: IARV(100p%|AIS≥n)=λn· 2.84−log(cid:0)1−p(cid:1) (6) IARV(cid:0)100p%|AIS≥1(cid:1)=β−(cid:113)αIn(cid:0)1−p(cid:1) (3) where,nisthespecifiedAISlevel,λnisthescaleparameterforthe specifiedAISlevel. UsingEq.6withtheassociatedscaleandshapeparameterval- where,αisthecutpointforthespecifiedAISlevel,βistheregres- ues,alongwiththeDARprobabilities,theIARVsinTable6can sioncoefficient.UsingEq.3withα=4.34andβ=671,alongwith bederived.Tosatisfythedesiredprobabilitiesofinjuryassociated the DAR probabilities, the IARVs for nominal and off-nominal withallfourAISlevels,theminimumBrICfornominalandoff- are340and470,respectively(Table6).Basedontheavailablelit- nominal conditions was selected. For nominal and off-nominal eratureanditsapplicabilitytotheenvironmentexpectedduring conditions,aBrICof 0.04and0.07,respectively,arerequiredso spaceflight,theconfidenceintheseIARVsisratedata4onascale thatallfourAISlevelsaresatisfied.Althoughthesevaluesaremost from0to5,with5beingthemostconfident. conservative,theBrICwasdevelopedforAIS≥4injurylevelsand thenscaledtootherAISlevels.Basedontheuncertaintyassoci- IARVforkinematicrotationalbraininjurycriteria ated with the extrapolation of the BrIC for low injury risk, the To further mitigate the risk of mTBI, an injury metric related confidenceintheseIARVsisratedata2. to angular head dynamics was selected. Takhounts et al. (2013) reportsthecalculationmethodforderivingbrainrotationalinjury IARVforneckaxialtension criteria(BrIC).TherevisedBrICcanbeappliedregardlessofthe TheTHORneckisdesignedtomimicthehumanneckresponse bymeetingbiofidelitygoalsbasedonperformancecorridorsfrom Mertzetal.(1997),NavalBiodynamicsLaboratory(NBDL),and JapanAutomotiveResearchInstitute(JARI)humanvolunteertests (Whiteetal.,1996).Althoughthegoalwastoproduceaneckthat isbiofidelicinallloadingconditions,theaxialstiffnessisgreater thandesired(Dibbetal.,2006).However,theaxialperformance oftheneckinloadingthroughthecenterofgravity(CG)canbe relatedtothecomputationmodelof PMHS.Thisisexpectedto betheprimaryloadingexpectedinNASAlandingsduetoinertial loads on the head. The computational model used was a vali- datedfiniteelement(FE)modelofthehumanneck.Notethatthe THORModKitATDwasrevisedtobehavesimilartotheTHOR- NT without muscle cables case, which correlates well with the computationalmodel.Usingthisrelationship,atransferfunction betweentheTHORModKitandthecomputationalmodelcanbe determined(Eq.9). THORaxialtensionrelationshiptoappliedforce: FIGURE2|Headinjurycriteria15injuryriskfunction(Funketal.,2007). FzTHOR =0.8228·Fapplied (7) FrontiersinBioengineeringandBiotechnology|Biomechanics March2014|Volume2|Article4|8 Somersetal. THORATDIARVsforspaceflight Table6|TableofproposedTHORinjuryassessmentreferencevalues(IARV). Conditioned Deconditioned IARVconfidencelevel(0–5) Nominal Off-nominal Nominal Off-nominal HIC15 340 470 340 470 4 BrIC 0.04 0.07 0.04 0.07 2 Neckaxialtensionforce(N) 880 1,000 760a 860a 4 Neckaxialcompressionforce(N) 580 1,100 500a 950a 3 Maxchestdeflection(mm) 25 32 25 32 2 Lateralshoulderforce(N) 2,700 3,300 2,700 3,300 4 Acetabularresultantforce(N) 1,600 2,900 1,200b 2,200b 3 Thoracicspineaxialcompressionforce(N) 5,800 6,500 5,000a 5,600a 3 Ankledorsiflexionmoment(Nm) 18 31 14b 23b 3 Ankleinversion/eversionmoment(Nm) 17 22 13b 17b 3 Averagedistalforearmspeed(m/s) 8.1 10 8.1 10 3 aSpinaldeconditioningfactorof0.86applied. bLowerextremitydeconditioningfactorof0.75applied. FIGURE3|BrainrotationalinjurycriteriaTHORinjuryriskfunctions (Saundersetal.,2012). FIGURE4|Neckaxialtensionforceriskfunctionsdevelopedfrom Philippensetal.(2011). PMHSaxialtensionrelationshiptoappliedforce: Necktensioninjuryrisk: F =0.5983·F (8) zPMHS applied p(AIS≥n|Fz)=1−Φ(cn−Fz ·β) (10) THORaxialtensionrelationshiptoPMHS: Necktensioninjuryriskmodel: FzTHOR =1.38·FPMHS (9) IARV(cid:0)AIS≥n|p(cid:1)= cn−Φ−1(cid:0)1−p(cid:1) (11) β AstudyreportedbyPhilippensetal.(2011)wasperformedto determinetheaxialtensionforceinjuryriskfunctionassociated where,nisthespecifiedAISlevel,ΦisthestandardGaussiandis- withPMHSinjuries.Usinganorderedprobitanalysisonthemax- tribution,CnisthecutpointforthespecifiedAISlevel,c1=6.30, imumreportedAIS(anatomicalandclinical),theriskfunctions c2=8.56,c3=9.28,c4=10.19,andβistheregressioncoefficient showninEqs10and11,andFigure4weredeveloped. (0.0053). www.frontiersin.org March2014|Volume2|Article4|9 Somersetal. THORATDIARVsforspaceflight UsingEqs9and11alongwiththeDARprobabilities,theIARVs as a starting point,the results of another Pintar et al. (1990a,b, inTable6canbederived.TheAIS≥1IARVs(880and1,000N 1998a)studywereinvestigated,whichwasbasedondatacollected fornominalandoff-nominalconditions,respectively)arelowest, previously.Equation12givesthe50%injuryriskfunctionbased andareselectedtobeconservative. onmultiplefactorsincludingloadingrate,age,andgender.Com- National Highway Traffic Safety Administration is proposing bining the logistic regression equation with Eq. 12 gives Eq. 13. anIARVof2,520Nforneckaxialtensionforautomotiveuseand Thisequationcanthenberewrittentodetermineinjuryriskbased relatestoa22%riskofanAIS≥3injury(Dibbetal.,2006).Based onaneckaxialcompressionforce(Eq.14). on the assumptions made and the comparison to the NHTSA Neckcompressionfracturetolerancemodel: IARVs,theconfidenceintheseIARVsisratedata4. Fz =β0+β1·A+β2·LR+β3·G+β4·A·LR (12) IARVforneckaxialcompression Neckcompressionisofparticularconcernduringspacecraftland- where,β0istheconstantcoefficientwithavalueof934.2,β1isthe ing, because there can be a significant +Z acceleration causing agecoefficientwithavalueof8.9,β2istheloadingratecoefficient neck compression from the inertial effects of the head. In addi- with a value of 11.0,β3 is the gender coefficient with a value of tion,anyhead-mountedmass,suchasaconformalhelmet,could 665.0,andβ4istheageandloadingratecoefficientwithavalueof increasetheloadontheneck. −0.134. Aswithnecktension,theassumptionwasmadethattheTHOR Neckcompressioninjuryrisk: neckisbiofidelic,andthuscadavericinjuryriskfunctionscanbe directlyapplied.Althoughthismaynotbethecase,itisassumed 1 p(AIS≥2|F )= (13) thatthemechanicalneckoftheTHORwouldbestifferthanthe z 1+eβ5(β0+β1·A+β2·LR+β3·G+β4·A·LR−Fz) human (similar to the trend seen in neck tension). Pintar et al. (1998a) conducted an experiment consisting of nine male and Neckcompressioninjuryriskmodel: four female cervical PMHS spines to developed neck compres- sion injury risk functions for axial compression. The mean age IARV(cid:0)AIS≥2|p(cid:1)=β +β ·A+β ·LR 0 1 2 of the subjects was 59years old (range of 39–82). These data (cid:16) (cid:17) log 1 −1 werecollectedduringhyperflexion,whichisthecauseof48–70% +β ·G+β ·A·LR+ p (14) of neck injuries (Allen et al., 1982; Yoganadan et al., 1989). As 3 4 β 5 this represents the most likely mechanism for injury and is the mostconservativeforneckcompression,thesedatawillbeusedto where,F isthepeakneckaxialcompression(N),pistheproba- Z developneckcompressionIARVs. bilityofneckinjury,β arethemodelcoefficients,Aisthesubject n Pintar et al. (1998b) report an injury risk function based on age(years),LRistheloadingrate(m/s),Gis0forfemales,and1 cervicalcompressionforce(Figure5A).Usingthisriskfunction formales. FIGURE5|Neckaxialcompressionforceinjuryriskfunction.(A)OriginalPintarriskfunction(Pintaretal.,1998a)withconfidencelimits(dashedlines)and (B)expandedriskfunctionsformaleandfemaleastronauts.Solidlinerepresentsmeanastronautage(47.1and43.3formalesandfemales,respectively), dashedlinesrepresenttheyoungestandoldestastronauts(male:35and56;female:32and54).Notethatforcetolerancedeclineswithage. FrontiersinBioengineeringandBiotechnology|Biomechanics March2014|Volume2|Article4|10