Flight Testing A Micro Air Vehicle Using Morphing F

FlightTestingAMicroAirVehicle

UsingMorphingForAerorvoelasticControl

MujahidAbdulrahim,HelenGarcia,GregoryF.Ivey,andRickLind

Microairvehiclesareabletoflyinenvironmentswithlittlemaneuveringroom;however,suchflightre-

quireshighagilityandprecisionmaneuvering.Acurrentclassofmicroairvehicleisconstructedwithonly

rudderandelevatorforcontrolauthority.Suchsurfacesarenotoptimalforflightcontrolbutaileronscan

notbeinstalledonthemembraneudforwingsonthevehicles.Morphingisudasanaerorvoelastic

effectorforcontrol.Vehiclesareconstructedusingthreadstocommandacurlofthewingsandusingtorque

rodstocommandatwistofthewings.Ineachca,flighttestsdemonstratetheactuationcaussufficient

deformationofthewingtoresultinsignificantcontrolauthority.Theflightdynamicsshowturnsandspins

canberepeatedlyperformedusingthisaerorvoelasticcontrol.

I.Introduction

Microairvehicles(MAV)arerapidlygainingattentionintheflighttestcommunity.Thesmallsizeandlightweight

ofthevehiclesmakethemespeciallyattractiveforsomemissions.Thevehiclesarequiteportablesomaybeeasily

transportedtoremotelocations.AMAVisinherentlystealthybecauofitssizeandquietbecauofitselectric

propulsion.Furthermore,manytypesofnsorsarebeingminiaturizedandwouldfitwithinthepayloadcapabilities

ofaMAV.

Someareasofoperation,suchasurbanenvironments,requireavehiclethatissmallbutalsoextremelyagile.

Agilityisobviouslyneededsothevehiclecanmaneuverwithinsmallareasandbetweenobstaclesinadenenviron-

ment.Thevehiclemustalsobehighlyresponsivetorejectdisturbanceslikewindgustswhichmaybeconsiderable

nearbuildings.

Theconceptofaerorvoelastictailoringthroughmorphingisbeingstudiedforcontrolonmanyvehicles.The

basicconceptofthisapproachinvolvesactivelychangingtheshapeofavehicletoaffecttheaerodynamicsand,con-

quently,theflightdynamics.Suchcontrolinherentlyusthecontrolandactuationthroughthestructuraldynamics

toaffecttheaerodynamicsasanaerorvoelasticsystem.Theresultingsystemhasdifficultdynamicssodesignand

analysisofaerorvoelasticcontrolisachallenge.

1

Controlthroughaerorvoelasticity,whilenotcommon,iscertainlynotanewconcept.Themostfamousexample

5

isarguablythewingtwistingudbytheWrightbrotherstooperatetheirflier.Academicstudieshavemainlyfocud

onfluttercontrolbutveralexamplesofmaneuveringdesignhaveincludedtheuofdynamicinversionandrobust-

2

nesstoaccountfortheaerorvoelasticdynamics.Inpractice,thedesignisoftenlimitedtosimplyformulatingnotch

3

filterstoavoidinstabilitiesbutretainlow-frequencyperformance.TheActiveAeroelasticWingprojecthasprobably

4

beenthefirstaircrafttodirectlyrequiredesignofanaerorvoelasticcontroller.

15

Morphinghasmorerecentlybeenconsideredaviablemethodofcontrol.Theapproachtochangetheshapeofa

6

vehiclewillclearlyaffecttheflightdynamicsandstaticpropertiessuchasliftanddrag.Issuesremaintobestudied

AmericanInstituteofAeronauticsandAstronautics

astotheuofmorphingfordynamiccontrol.Forinstance,studiesneedtobeperformedascertainingifavehiclebe

morphedwithsufficientmagnitudeandbandwidthtocontroltheshortperiodmotion.Inthisca,theuofmorphing

asaerorvoelasticcontrolforagilityistheissue.

Thispaperconsiderstheuofmorphingasanaerorvoelasticeffectorforcontrolofamicroairvehicle.Several

typesofmorphingareactuallyconsideredonveraltypesofvehicles.Avehiclewith12inwingspanhasthreadsthat

curlthewings.Anothervehiclewith24inwingspanhastorquerodsthattwistthewings.Ineachca,themorphing

isshowntogeneratesignificantcontrolauthorityformaneuvering.Propertiesofturnsandspinsarehighlightedto

demonstratetheeffectofthisaerorvoelasticcontrol.

II.Morphing

Theconceptofmorphingisafairlybroad-rangingidea.Amorphingaircraftisgenerallyacceptedtobeanaircraft

whoshapechangesduringflighttooptimizeperformance.Suchchangesmightincludespan,chord,camber,area,

6

thickness,aspectratio,planformandanyothermetricrelatedtoshape.Themorphingcanevenbeappliedtoacontrol

surfacetoeliminatehinges.

29

Thecontinuingmaturationofmaterialsandcontrolstechnologyisleadingtoconsiderationofmoreextremetypes

ofmorphingforcontrol.Aircraftcanbemorphedinveralbiologically-inspiredwaysthatareappropriateforcontrol.

Forinstance,birdshaveverycomplexwingshapesduringflightwithlargevariationsinsweepandtwistalongthe

entirespan.Furthermore,thespanitlfchangesbadonflightconditionformanybirds.Thecomplicatedmorphing

isactuatedusingjoints,suchaselbowandwrist,alongwiththefeatherstoachievemanywingshapes.

ProgramsbyDARPAandNASAhavebeenextensivelystudyingtheconceptofmorphing.Thestudieshave

showntheaerodynamicbenefits;however,theuofmorphingforcontroldesignhasnotbeenstudiedextensively.

Thelimitedrearchintocontrolhasactuallyconsideredstaticissues.Anaircraftwithmorphinghasbeenshown

tohavegoodperformancewhenfixedatdifferentconfigurations.Controlwasconsideredformorphinginarelated

21

studybutonlyconsideredtheamountofactuationenergyneededtocaumorphing.Theamountofcontrolauthority

26

wasalsostudiedforsmartmaterialsembeddedinastructureinthecontextofsteady-stateopen-loopcontrolledflight.

23

Theissueofcontroldesignformorphingsystemshaspartlybeenoverlookedbecauofthelackofexperimental

testbeds.Manymechanismsformorphinghavebeenconstructedbuttheyarenotyetimplementedonaflightvehicle.

NASAhasconstructedawingwithmorphingcamber.Atofsmartsparshasbeenbuiltthatprovidemanytypes

24

ofmorphing.Mechanismshavealsobeenbuiltsuchasatelescopingspartomorphaspectratioandaninflatable

1920

actuatortochangesweep.Ineachca,themechanismsaretooheavyforaflightsystemsoonlywindtunneltesting

28

ofastaticmounthasbeenperformed.Aninflatableconcepthasactuallybeentakentoflightbutthemorphingisa

22

singleone-timeinflationsoisnotudformaneuveringcontrol.

AprojectparticularlyrelevanttothestudyofmorphingforcontrolistheF/A-18ActiveAeroelasticWing(AAW).

15

TheAAWisusingwingtwisttoaltertheaerorvoelasticdynamicsandgeneraterollmoments.Themorphingisac-

tuallyapassiveeffectresultingfrommovingleading-edgesurfacesoftheflexiblewing.Thattestbedudasimple

mechanismtogeneratemorphinginordertostudytheresultingloadsandcontrolissues.

AMAVisanespeciallyappropriateplatformforconsiderationofmorphing.Theflexibilityofthemembrane

enablestheshapeofthewingstobeeasilychanged.Theresultingchangewillobviouslyaffectthedynamicsand

actasaneffectortoprovidecontrolauthority.Assuch,themorphingofaMAVaddsverylittlecostbutprovides

tremendousbenefit.Thispaperwillconsidersimplestrategiestoprovidemorphingtoemphasizecontrolissues.

Esntially,themethodsofmorphingarenotconsideredasmuchasthedynamicsofmorphing.

III.VehicleDesign

Small,remotelypilotedvehiclesareadaptedformorphingthroughsimplemodificationofthestructure.The

vehiclesarebetween10inand24ininwingspananduavarietyofstructuralmemberstobothwithstandflight

loadsandaccomplishmorphingdeformation.Vehiclesofthissizeareespeciallywellsuitedastestbedsforsimple

morphingstrategiesbecauofthehighstrengthoftheactuatorscomparedtoflightloads.LightweightMAVairframes

2of17

aresubjectedtorelativelysmallairloadsduringflightandmaneuvering.Thesmallaerodynamicloadspermitflexible

materialstobeudthroughoutthestructure.Suchflexiblestructuresareeasilydeformedusingexistingactuators.In

turn,simplemorphingstrategiesforsmallairvehiclesarereadilydevelopedandflighttested.

A.WingCurlingAircraft

Morphingcontrolwasfirstimplementedona10inwingspanmicroairvehicle.Thisvehicleudelevatorand

ruddersurfacesforcontrol,bothofwhichweremountedonaconventionalempennage.Afolding,flexiblewingwas

attachedtothefulageofeachaircraft.Thewingwasmountedoncabanestrutsforthe10inwingspanvehicle,

andonmid-fulagerodsonthelarger12invehicle.Thedifferentwingpositionallowedeachvehicletoundergoa

slightlydifferenttypeofmorphingdeformation.Thetwovehicles,showninFigure1,areesntiallysimilar,apart

fromthewingpositionandtheinclusionofruddercontrolonthe10invehicle.Otherwi,bothvehiclesaccomplish

morphingbycurlingthecompositewingsandchangingtheangleofincidenceononewingatatime.Turncontrol

isaccomplishedbymorphingasinglewingtoincreatheaerodynamicforcesandgeneratearollmomentinthe

directionoppositetothemorphedwing.Forinstance.curlingtheleftwingresultsinapositive(right-handed)rollrate

andanensuingrightturn.

Figure1.WingcurlingmorphingMAVs-10inwingspanhigh-wingaircraft(left)and12inspanmid-wingaircraft(right)

Thefolding,flexiblewingusastripofcarbonfiberweaveastheprimarystructuralmember.Thestripextends

approximately1inbackfromtheleadingedgeofthewing,reachingthepointofmaximumcamber.Thinstripsof

unidirectionalcarbonfiberareembeddedwithintheleadingedgestripandextendtothetrailingedge.Theentire

structureiscuredagainstafemalemoldtoformtheairfoilshape.Finally,anextensiblelatexmembraneisadheredto

theleadingedgeandbattens.Theresultisathin,undercamberedwingsurfacehavingbothflexibilityinthebattens

andcurlingcapabilityintheleadingedgestrip.Thecurvedleadingedgectionisabletoresistconsiderablebending

loadsinthepositivedirection.Undernegativeloading,however,thestructurereadilybucklesanddeformsinboth

curlingandtwisting.

ThistypeofwingstructureisnormallyudforcollapsingthewingofaMAVtofitintoasmallcontainer.It

isalsowellsuitedformorphing,giventhecomplianceofthestructure.Themorphingiscontrolledbyusingsmall

rotaryactuatorsconnectedtohardpointsonthewingstructurebytensionedKevlarcables(Figure2).Astheactuator

adjuststhetensiononthecable,thewingdeformsintoatwistedformthatisappropriateforflightcontrol.Namely,the

resultingshapeincreastheangleofincidenceofthemorphedwingandincreastheliftingforceproduced.When

onewingsideismorphed,aliftdifferentialiscreated,causingtheaircrafttoincurarollrate.

TheextentandshapeofthemorphingcanbeadjustedbyvaryingtheamountoftensionintheKevlarlinesor

adjustingthelocationofthehardpointonthewing.Theshapeisalsodependentonthedirectionofthetensileforce

fromtheKevlar,whichisdeterminedbythepositionoftheactuatorarmwithrespecttothewinghardpoint.Alarge

verticalparationbetweenthetwopoints,asonthe10inaircraft,producesalargelytwistingmotion.However,

asthetensileforceisappliedinamorespanwidirection,asonthelowerwing12inaircraft,thewingexhibitsa

predominantlycurledmotionduringactuation(Figure3).

3of17

Figure2.Undersideviewoffolding,flexiblewingwithtensionedKevlarcables

Figure3.Rearviewof12”MAVshowingundeflected(left)andmorphedwing(right)

Withouttwisting(andstrictlycurling),thewingmorphingdoesnotaffecttherolltrimtosufficientlycontrolthe

aircraft.The12inairplanerequiresanadditionalKevlarstrandconnectedtoahardpointonthetrailingedgetoproduce

asufficientamountofwingtwisting.Esntially,asingleactuatoroutputarmisudtotensiontwocablesconnected

totwopointsonthewing.Theresultingmorphingproducesashapedeformationthatissuitableforlateral-directional

control.Theshapeandthicknessofthecarbonfiberleadingedgestriphasaconsiderableeffectonthemorphing

shape.Thewingsaresomewhatnsitivetodesignorfabricationchanges,whereadisparityinthestructurecancau

adrasticallydifferentmorphingshape.

B.WingTwistingAircraft

Theconceptoftakinganexistingdesignandretrofittingamorphingmechanismwasextendedtoa24invehicleof

similargeometry(Figure4).Aswiththesmalleraircraft,thevehiclewasbadonconventionalconfigurationandud

anelevatorandrudderforcontrol.Theprimarydifferencebetweenthetwoaircraftisthewingstructure,whichus

anominallyrigidunidirectionalcarbonfiberleadingedgeinsteadofthemorecompliantcarbonweave.Additionally,

thewingsurfaceisalessextensiblenylonfilm.Boththewingsurfaceandleadingedgestructurereducethemagnitude

ofthedeformationthatresultsfromflightloads.However,thewingretainedtheflexibilityinthebattens,whichallows

forwingwarpingviaanaluminumtorque-rod.

4of17

PropertyWingTwistingMAVWingCurlingMAV

AmericanInstituteofAeronauticsandAstronautics

Theprimarydifferenceinconceptbetweenthe24inwingwarpingandthe12inwingcurlingisthebidirectional,

positivecontroloftheshapedeformation.Thetorque-rodsareabletoactuatethewingintwistinbothtrailing-edge

upandtrailing-edgedowndirections.Suchcontrolisnotpossibleusingthetensionedcables,asthemechanism

reliesonthewingstructuretoprogressivelybuckleandcurl.Shouldthewingleadingedgedeformtoosoonortoo

far,thecablewoulddevelopslackanddeterioratecontrolofthewingshape.Sincethewingwarpingaircraftusa

stifferleadingedgematerial,theproblemofprematurebucklingorexcessivedeformationunderaerodynamicloadis

largelyeliminated.Thus,thecontrolofthewingshapeislargelyafunctionoftheactuatorposition,althoughthewing

remainesfreetodeformslightlyinrespontoairloads.

IV.FlightTestingandHandlingQualities

A.TurnsandBasicControllability

1.WingCurling

Thecontrollabilityofthewingcurlingaircraftisimprovedoverarudder-elevatorequippedaircraft.Thewingmor-

phingfacilitatesflightpathtrackingoverthemajorityoftheflightenvelope.Atcruiairspeed,themorphingcontrol

issufficientforsmall,stabilizingadjustmentsabouttriminadditiontolargecommands.Althoughsomeroll-yaw

couplingresultsfromthemorphingdeflection,theresponremainsgreatlyimprovedoverruddercontrolforturns.

Thewingcurlingmorphingexhibitsgoodcontrolresponneartheneutral,trimposition.Smallinputsareneces-

saryinperformingturnsandinmakingslightadjustmentstotheflightpath.Underthecircumstances,themorphing

providesanadequatelevelofcontrol.Althoughthephysicaldeformationofthewingsurfaceisnotnecessarilylin-

ear,theaircraftrespondspredictablytovariousmagnitudesofcontrolinput.Inparticular,themorphingissuitable

forbothcommandingturnsandforcorrectingforattitudeperturbationsfromwindgustsorotherdisturbances.Roll

controllabilityremainssatisfactorythroughouttheairspeedrangeencounteredduringcrui,high-speeddives,and

landing/approachphas.

However,rollhandlingqualitiestendstobequitensitivetothelocationofthehardpointonthewingandtothe

tensioninthecable.Slightasymmetriesintherightandleftsidecabletensionsoftencontributestodifficultiesin

controlandnon-zerotrimcondition.Overariesofflights,thecontrolresponchangedslightlyfromvariationsin

thecontrollinkages.Additionally,deteriorationofthelatexmembranenoticeablyreducedthewingsurfacetension.

Thenaturalrubberudinthelatexmaterialdecayedwhenexpodtothesun.Asaresult,thereducedtension

preventedthedeformationfrompropagatingsmoothlythroughoutthewingstructure.Inturn,thetwistdeformation

caudbythebucklingremainedlocalizedaroundthehardpointandreducedcontroleffectiveness.

2.WingTwisting

Turnperformanceisconsiderablyimprovedwiththewingtwistmorphingonthe24inaircraft.Thetorque-rodmech-

anismfacilitatesbasictaskssuchascommandingabankangleandcorrectingforattitudeperturbations.Theaircraft

respontosmallcommandinputsismoreimmediateandconsistentthanthewingcurlingaircraft.Thisimprovement

occurredinpartduetotheimprovedcontrolpoweroftheactuatoroverthemorphingdeflection.Additionally,the

wingtwistgeneratesananti-symmetriccommandthatmorphsthewinginbothpositiveandnegativedirectiononboth

wingssimultaneously.

Inadditiontosubstantiallyimprovingcontrolpower,theanti-symmetricmorphingreducesyawcouplingcompared

tothewingcurlingmorphing.Minimalruddercorrectionsarerequiredtomaintainturncoordinationduringmaneu-

vering.Theimprovedmorphingmechanismofthe24inaircraftcontributedtovastlyimprovedhandlingqualities.

Inparticular,theaircraftperformanceingustywindconditionsbenefitsfromthepositivecontroloverthemorphing.

Highfrequencycontroltaskssuchasattitudecorrectionandstabilizingcontrolareeasilyperformed.

6of17

B.360Rolls

Continuous360

rollmaneuversareperformedforeachaircrafttodeterminethemaximumrollrateachievablewith

morphing.Thetestshelpdeterminetheresponoftheaircrafttoalongmorphingcommand.Theyarealsouful

inidentifyinganyobviouscross-couplinginpitchandyawthatoccurduringthemaneuver.Rolltestsareperformed

withtheaircraftattrimmed,straightandlevelflightatcruiairspeed.Fullmorphingiscommandedtoeitherrightor

leftdirectionsandhelduntiltheaircrafthascompleteda360

rotationalongtherollaxis.

1.WingCurling

Thewingcurlingmorphingprovidesasufficientlevelofcontroltoeffectacompleteroll.Despitetheasymmetry,

theaircraftexperiencesarelativelysmalldivergenceinflightpathduringthemaneuver.Undermaximummorphing

deflection,astabilizedrollrateisachievedwithin0.5conds.Maximumrollrateisapproximately,achievable

inbothleftandrightrolldirections.

Duringtherollmaneuver,theaircraftprescribesanalmosthelicalflightpath.Aslightgaininaltitudeisincurred

halfwaythroughthemaneuver,followedbyadescendingreturntolevelflight.Atthecompletionoftheroll,the

aircraftistypicallypitcheddownslightlyandmayhaveasmallheadingchange.

2.WingTwisting

Theaircraftisabletoachieveahighlevelofrollperformanceusingwingtwistmorphing.Theperformanceis

improvedcomparedtowingcurling,rudder,orevenconventionalailerons.Atthemaximummorphingdeflectionof

approximately10twist,theaircraftachievesastabilizedrollrateofwithin0.25conds.Atnearly3rolls

percond,aircraftexperiencesverylittleflightpathdivergenceorcross-coupling.Duringthemaneuver,theaircraft

wingsbecomevisiblyblurredfromtherotation,whilethefulageappearstorotateaboutthecenterlineaxis.Similar

MAVsusingrigidwingsandhingedconventionalaileronswereabletoachieveamaximummeasuredrollrateof

.

Recoveryfromtherollmaneuverisalsowithin0.25condsofneutralizingmorphingcommand.Rollperformance

changesslightlyovertheairspeedrange,withhigherairspeedsexhibitinghigherrollrates.However,thecontrol

remainspositiveandsufficientfromnear-stallspeeduptothemaximumdivespeed.

C.Spins

Controlofamorphingvehiclebeyondthestallboundariesisanotherrelevantfacetoftheflightdynamics.Agreater

degreeofcontroloverthevehiclegeometrymayimprovestall/spinavoidanceor,converly,evencommandade-

velopedspinmaneuver.Thelargedegreeofcontrolaffordedbythemorphingmechanismcouldbebeneficialin

generatinganti-spinforcesandrecoveringfromastabilizedspin.

Thespincharacteristicsofthevehiclesareinvestigatedbymanually-pilotedmaneuvers.Spinmodesareiden-

tifiedusingclassicalspinentrytechniques.Theindividualmodesareproducedbytrial-and-errorandhavebeen

18

reproducedoververalflighttests.Thefollowingdiscussionprentssomeofthepredominantspintypesencoun-

teredduringtheflighttests.Notethespinsallumorphingbecauthespincharacteristicswerenotnearlyas

significantusingonlyrudderandelevator.

1.WingCurling

Spinmodesofthe10inand12invehicleswereinitiallyencounteredduringflighttestsofaggressivemaneuvers.

Largemorphingcommandscoupledwithpositiveelevatorcommandsresultedinsuddenandviolentrollsoppositeto

thecommandeddirection.Thisbehaviorwasmostoftenobrvedatlowerairspeeds,leadingtotheidentificationofa

considerabledeficiencyinthewingcurlingmethod.

Thewingcurlingmorphingisabletocontrolrollrateprimarilybyincreasingtheangleofattackofthemorphed

wing.Thedeformationalsoencompassspanwicurling,althoughtheeffectoftwistingtoincreatheincidence

7of17

appearstobethepredominantfactorcontrollingrollmoment.Undertypicalflyingconditionsandmoderatemorphing

deflections,themorphingisabletogenerateaproverliftdifferential(i.e.inthecommandeddirection)andeffecta

rollratechange.Thisisthesamecharacteristicdescribedintheturnandrollperformancections.

However,asthemorphingcommandbecomeslarge(causingalargechangeinincidenceangle)andiscoupled

withpositiveelevatordeflection,theangleofattackofthemorphedwingcanexceedthestallangle,causingavere

lossofliftinthedeformedregion.Theresultingliftdifferentialbetweenthetwowingsbecomesadvertothe

commandeddirection,leadingtoanopposite-directionroll.Whilethisisoccasionallyrecoverable,suchanasymmetric

stalltypicallydivergesintoaspin.Theoverwhelmingmajorityoftheencounteredspinswereterminal,resultingin

theaircraftimpactingthegroundinastabilizedspin.

Recoveryfromastall-spinisquitedifficultforanumberofreasons.Therearcherssuspectthatthestallresulting

fromexcessivemorphingpositioncausacollapofthelow-pressureregionabovethewing.Thiscollapreduces

thetensiononthecontrolcableandpreventsthewingfromproperlyextendingandreturningtotrimflightcondition.

Additionally,themorphingmechanismpermitsthepilottostrictlyincreatheangleofattackofthewings.Thus,

inopposingtherollrateincurredduringaspin,thepilotmustincreatheangleofattackoftheopposite(unstalled)

wing,whichmayfurtherdecaythecontrollabilityoftheaircraft.

Somespinstestsresultedinasuccessfulrecoverytolevelflight.Thespinsweretypicallyinitiatedusingagentler

stallentryandsmallermagnitudesofelevatorandmorphingcommandinputs.Flightpathduringthespinsrembled

adescendinghelicalspiral.Recoveryprocedureentailedneutralizingcontrolsurfaceandmorphingdeflectionand

allowingtheMAVtolf-recoverfromthespin.Oncestabilized,elevatorcommandwasudtopullupfroma

verticaldivetolevelflight.However,becauofthedifficultiesinconsistentlyrecoveringfromadeparture,spins

werenotconsideredasufulmaneuversforthistypeofMAV.

2.WingTwisting

Figure6showsthecommandandrotationratesduringaconventionalspin.Thismaneuverisinitiatedfromlevelflight

bycommandingpositiveelevatortoincreathepitchrateandangleofattack.Rightruddercommandisthenapplied

togenerateayawingmomentastheaircraftapproachesstall.Inthisca,theyawcausanasymmetricstalland

startsthespinrotation.Theaircraftresponisrelativelyconstantthroughoutthemaneuver,althoughtherollrate

tendstobuildupastheflightpathchangesfromleveltoavertical.Theautorotationcontinuesaslongasthepositive

elevatorandruddercommandsareheld.Oncethecommandsareneutralized,therotationslowsandcomestoastop

withlittleornooppositerudderinput.Positiveelevatorisudtorecovertheaircrafttolevelflightat363conds.

Althoughthistypeofspinhasbeenexperiencedveraltimes,theentryprocedurestendtobedifficulttoreproduce.

Specifically,applyingruddercommandatalowangleofattack(tooearly)preventsastallfromdevelopingandresults

inahigh-speedspiraldive.BothwindtunnelandCFDanalysishaveshownthatthethin-undercamberedairfoilsud

onthevehiclehavedelayedstallrespon.Thisaffordssuchvehiclesincreadresistancetostall-spindeparture,at

leastforpositiveloadings.

Theeffectofmorphingonpositive(upright)spinsistoacceleratetheontofthespinandtoassistintherecovery

process.Thiseffectismostpronouncedduringcross-coupledcontrols,wheretherudderdirectionisoppositetothat

ofthemorphing.Insuchaca,thehighangleofattackattheinsidewingtipisfurtherincreadbythemorphing

actuation,leadingtoanobrvedstall-spin.Releasingthemorphingcommandeffectivelyreducesthewingangleof

attackandproducesnearlyimmediaterecoveryfromanupright,conventionalspin.

Conventionalspinsarealsoperformedwithnegative(down)elevatoractuationtoproduceastarklydifferentre-

spon.Inparticular,thespinmodesobrvedareofconsiderablyhigherenergy.Therotationratesofanegative

spincomparedwithanuprightspintendtobebetween2to6timesgreater.Badonrudimentaryanalysis,thestall

characteristicsofathinunder-camberedwingatnegativeanglesofattackarefarmoreverethanthecharacteristics

athighanglesofattack.Inflight,theairplaneisobrvedtohaveaveryimmediateandviolentrespontolarge

negativeelevatorcommands.Suchaninputisbelievedtocauanegativestallquickly,whereanyasymmetryabout

theyawaxisproducesalargerateofrotation.

Figure7showsanidentifiednegativespinmodeinitiatedbyamorphingcommandwithelevatorandrudder.At

401conds,theaircraftrespondstotheconstantcontroldeflectionbybuildinguprotationratesonallthreeaxis.The

8of17

40

30

50

10

0

R

e

s

p

o

n

s

e

(

d

e

g

/

s

)

C

o

m

m

a

n

d

(

d

e

g

)

20

0

−10

−20

−30

−40

359360361362364363

elevator

rudder

morphing

−50

roll rate

pitch rate

yaw rate

360361359362363364

−100

Time (s)

Time (s)

Figure6.PilotCommands(left)andRespons(right)duringconventionalspin

entryintothemaneuverisrelativelygradualandonlyafteronecondofcontrolinputshavethepitch,roll,andyaw

ratesbecomesignificant.

Thisparticulartypeofspintendstostabilizeindependentlyoftheinitialpro-spincontroldeflections.Att=402

conds,thecontrolsarerelead,whiletheaircraftcontinuestospin.Theapplicationofpositiveelevator(forrecov-

ery)shortlyafterwardsappearstomaintainthespinforsometime.Itisonlywithcorrectiveoppositeruddercommand

thattheaircraftarreststherotationandrecoversfromthespin.

Itisdifficulttodrawsolidconclusionsfromthisspinquence.However,therearchersattributethetwodistinct

modesobrvedtobeacaofprimaryandcondaryspincharacteristics,wherethelatteriscaudbyapremature

recoveryattempt.Similarspinshavebeenobrvedfrominbothleftandrightdirections.

40

30

elevator

rudder

morphing

50

10

0

R

e

s

p

o

n

s

e

(

d

e

g

/

s

)

C

o

m

m

a

n

d

(

d

e

g

)

20

0

−10

−20

−30

−40

400401402403404405

−50

roll rate

pitch rate

yaw rate

401402400403404405

−100

Time (s)

Time (s)

Figure7.PilotCommands(left)andRespons(right)duringspin

Alternatively,Figure8showsaconsiderablydifferentspinbehavior.Althoughinitiatedbycommandssimilarto

thepreviousspins,thistypeofspinexhibitsacyclicorperiodicmotion.Itisperhapswiththetimingofthecontrol

inputsthatadifferencecanbefound.WhereasinFigure9,theelevatorinputlaggedbehindtherudderandmorphing

inputs,thespindepictedbyFigure8showstheelevatorleadingslightly.Theprecieffectthishasontheairflowis

unknown.However,theresultingaircraftresponisshowntobe6timesgreaterinmagnitudethanaconventional

spin.

Fromlevel,trimmedflight,theaircraftissubjectedtofullleftwingmorphing,fullleftrudder,andfullnegative

elevatorcommand.Theinitialreactionoftheaircraftistopitchdownataconstantrateandincuraleftrolland

9of17

yawfromthewingmorphingandrudderdeflections.Oncethewinghasreachedthenegativestallangle,presumably

facilitatedbythedeflectedwing,arapidspinensues,nearlydoublingtherollandyawratesandreducingpitchrate.

Thispatternisrepeatedfourtimesthroughoutthespin,allwhilepilotcommandsareheldconstant.Eachcycleis

proceededbyaperiodoflowmomentum,followedbyasharpchangeinpitchratealongwithpeaksinboththeroll

andyawrates.

Whilethedynamicsofsuchamaneuverarenotverywellunderstood,itappearsthatthemorphingofthewing

playsalargerollinbothinducingandrecoveringfromthespin.Forinstance,similarspinentriesperformedwithout

morphingarecharacterizedbyconsiderablylowerrotationratesandacontinuationofthespinaftercommandinputs

areneutralized.However,therecoveryofthiscyclicspinmodeoccursnearlyimmediatelyafterthecontrolsare

neutralized.Asenatt=176inFigure8,theaircraftisattheperiodofhighestmomentduringreturntoneutral

command.Therotationratescontinuetofollowthecharacteristicspikepatternandfinallyconvergestozerorotation

rates.

40

30

elevator

rudder

morphing

50

roll rate

pitch rate

yaw rate

10

0

R

e

s

p

o

n

s

e

(

d

e

g

/

s

)

C

o

m

m

a

n

d

(

d

e

g

)

20

0

−10

−20

−30

−40

171172173174175176177

−50

−100

171172173174175176177

Time (s)

Time (s)

Figure8.PilotCommands(left)andRespons(right)duringcyclicspin

Inflight,thishastheeffectofstoppingtheaircraftinmid-rotation.Unliketheotherspinmodesobrved,the

cyclicspinmodehasnoapparentrecoveryapartfromneutralizingthecontrols.Theaircraftwillcontinuetotheendof

agivencycle,cearotation,andsimplyflyaway.Theno-downrecoverytypicalofotherspinmodesiscontrasted

withanimmediaterecoverytolevelflight.

TheufulnessofthecyclicspinmodedepictedinFig.8isperhapsquestionable,althoughitmaygiveritoa

differentmodeofmaneuveringformorphingaircraft.Forinstance,theabovemaneuvermaybeufulforacontrolled

verticaldisplacement.Oninitiatingtheentry,theairspeedquicklydecaysandstartstheaircraftonarelativelyslow

verticalflightpath.Duringthisportionofthemaneuver,theaircraftincursariesofhighrateofrotations,each

paratedbyaperiodoflowmomentum.Asevidencedbytherecoveryfromthemaneuver,thisperiodcanbeud

torecovertheaircraftintostableflight.Whilepreviousspinmodesrequiredcorrectiverudderandsignificantaltitude

lossforrecovery,thiscyclicspinmodestoppedoncethecontrolswereneutralized.

Attitudeandairspeedentryconditionsintothespintrialshavebeenobrvedtohavesomeimpactonthestabilized

spinmodes;however,accuratemeasurementsoftheentryconditionswerenotpossible.Thelackofpressurensors

ontheairframeprecludedthegatheringofsuchdata.Excitationofaparticularspinmodedependedonthepilotability

topositiontheaircraftproperlybadoncontrolfeelandvehicleobrvations.

Thespinentrymaneuverswerealsoattemptedforothercontrolcombinations.Specifically,cyclicspinswere

attemptedwithoutwingtwistingbyusingnegativeelevatorandrudderdeflection.Thetrialsresultedinastabilized

spinbutwithconsiderablylowerrotationratesthanthecyclicspin.Additionally,thismodedidnotexhibittheperiodic

behaviorachievedthroughwingtwistingduringaspin.

10of17

V.Modeling

A.WingCurling

Flightdatafromthevehicleisanalyzedtoestimatemodelsoftheflightdynamics.Severaltechniqueswereattempted

toestimatethemodels,includingsystemidentificationandparameterestimation,butwithlimitedsuccess.This

2732

vehicleisparticularlydifficulttomodelbecauthemorphingcaustime-varyingasymmetrieswhichviolatemany

assumptionsudbystandardroutines.

Furthermore,theestimationisdifficultbecauoflimitedflightdata.TheMAVisequippedwithonlygyrosand

accelerometersbuttheflightdatafromtheaccelerometersisactuallytoonoisytobeufulformodeling.Thus,veral

criticalmeasurements,suchasangleofattackandangleofsideslip,arenotavailable.Somedynamicsarenoteasily

obrvable,especiallyintheprenceofnoi,usingonlytheavailablensors.

Anonlinearauto-regressivemodelisudtoreprenttheflightdynamics.Thegeneralformofthismodelis

showninEquation1.Thismodelrelatesthegyromeasuresofrollrate,

,pitchrate,,androllrate,,tothe

morphingcommand,

,andelevatorcommand,,atthesamplinginstanceof.Thematrices,

and

,reprentthedynamics.

(1)

ThemodelinEquation1containsquadratictermsoftheratesandcommands.Suchquadratictermsareincluded

toaccountforunknownrelationshipsbetweenthewingshapeandtheaerodynamics.Inthisca,thetermsutilizean

absolutevaluetoallowthecontributionsfromthequadraticstochangeinsign.

ThemodelinEquation1alsocontainscouplingterms.Thetermsmultiplythegyromeasurementsbyeach

other.Thestandardequationsofmotionforarigid-bodyaircraftincludecouplingtermswhichscalebythemoments

ofinertia.ThisMAVisobviouslyasymmetricduringthemorphingsothecouplingareesntial.

33

Finally,Equation1computestheupdatetothegyromeasurementsasafunctionofthemeasurementsfromtwo

previoussamplingtimes.Thetermsareincludedtoaccountforthetime-varyingnatureofthedynamicswhichari

byalteringthewingshape.Thedynamicsareassumedtobesufficientlydescribedbytwosamplingtimesalthougha

rigorousstudyofthesamplingtimeswasnotconducted.

Thevaluesofthematrices,

and,inEquation1aredeterminedbyaleast-squaresfittotheflightdata.The

resultingmodelisudtosimulatetheresponstothemorphingandelevatorcommands.Suchresponsareshown

inFigure9.

TheresponsinFigure9demonstratethemodelcapturesthebasictrendofthedynamicsbutisnotcompletely

accurate.Thepredictedresponsarenotperfectmatchestothemeasuredresponsbutyettheyclearlyshowsimi-

larities.Thus,themodelindicatesthetime-varyingasymmetriesassociatedwiththemorphingcausnonlinearities

andcouplingintheflightdynamicsofthisMAV.

11of17

155

10

datadata

simsim

8

6

data

sim

50

0

−5−5

−10

−15−10

01234560123456

P

i

t

c

h

R

a

t

e

(

d

e

g

/

s

)

R

o

l

l

R

a

t

e

(

d

e

g

/

s

)

Y

a

w

R

a

t

e

(

d

e

g

/

s

)

4

2

0

−2

−4

−6

Time (s)Time (s)

−8

0123456

Time (s)

Figure9.MeasuredandPredictedResponsforRollRate(left),PitchRate(middle)andYawRate(right)

B.WingTwisting

Flighttestingoftheactivewing-shaping24inMAVisperformedintheopenareaofaradiocontrolled(R/C)model

fieldduringwhichwindconditionsrangefromcalmto7knotsthroughouttheflights.Oncetheflightcontroland

instrumentationsystemsarepoweredandinitialized,theMAVishand-launchedintothewind.Thislaunchisan

effectivemethodtoquicklyandreliablyallowtheMAVtoreachflyingspeedandbeginaclimbtoaltitude.

ThisairplaneiscontrolledbyapilotonthegroundwhomaneuverstheairplanevisuallybyoperatinganR/C

transmitter.Thedataacquisitionsystembeginsrecordingassoonasthemotorispowered.

Thisaircraftdesignallowseitherrudderorwingshapingtobeudastheprimarylateralcontrolforstandard

maneuvering.Theairplaneiscontrolledinthismannerthroughturns,climbs,andlevelflightuntilasuitablealtitude

isreached.Ataltitude,theairplaneistrimmedforstraightandlevelflight.Thistrimestablishesaneutralreference

pointforallthecontrolsurfacesandfacilitatesperformingflighttestmaneuvers.

Open-loopdataistakentoindicatetheflightcharacteristicsoftheMAV.Specifically,theratesandaccelerations

aremeasuredinrespontodoubletscommandedparatelytothervos.Severaltsofdoubletsarecommanded

ranginginmagnitudeanddurationtoobtainarichtofflightdata.

ThedynamicsoftheMAVinrespontoruddercommandsisinvestigatedtoindicatetheperformanceofthe

traditionalconfigurationforthisMAV.Areprentativedoubletcommandandtheresultingaircraftresponsare

showninFigure10.

15

10

150

100

50

0

−50

−100

−150

150

100

Y

a

w

R

a

t

e

(

d

e

g

/

s

e

c

)

123450

R

o

l

l

R

a

t

e

(

d

e

g

/

s

e

c

)

R

u

d

d

e

r

C

o

m

m

a

n

d

5

0

−5

−10

−15

0123456

50

0

−50

−100

−150

−200

012345

Time(c)

−200

Time(c)

Time(c)

Figure10.DoubletCommandtoRudder(left),RollRaterespon(middle),andYawRaterespon(right)

TherollrateandyawratemeasuredinrespontothiscommandareshowninFigure10.Therollrateissufficiently

largeandindicatestherudderisabletoprovidelateral-directionalauthority;however,theyawrateisclearlylarger

thandesired.Actually,theyawrateissimilarinmagnitudetotherollratesothelateral-directionaldynamicsarevery

tightlycoupled.Theeffectoftherudderinexcitingthedutchrolldynamicsisclearlyevidencedinthisrespon.

12of17

Doublets,suchasthepulquenceshowninFigure11,arealsocommandedtothemorphingrvo.

8

6

150150

100100

5050

00

−50−50

−100−100

−150−150

M

o

r

p

h

i

n

g

C

o

m

m

a

n

d

4

2

0

−2

−4

−6

−8

00.511.52

Y

a

w

R

a

t

e

(

d

e

g

/

s

e

c

)

0.511.5200.511.520

R

o

l