Conjoined Lander-Rovers for Planetary Exploration

Steven A. Huber CMU-RI-TR-09-11

April 2009

Robotics Institute Carnegie Mellon University

Pittsburgh, Pennsylvania 15213

CarnegieMellonUniversity

Submittedinpartialfulfillmentoftherequirements

forthedegreeofMasterofScience

2

AbstractThis research explores the configuration of conjoined lander‐rovers for planetaryexploration.Technicalevolutionhasrefined landersandrovers,and improvementhas been continuous, but ongoing evolution offers no leap of performance for agiven launch. As long as rovers and landers are distinct, there is inevitableredundancy of structure, power, computing, and communication. For traditionalroversandlanders,eachredundantcomponent ismanifestedinmassandvolume,which precludes minimization of mobile planetary exploration systems. Theopportunitydevisedistoeliminateboundariesdemarcatinglanderandrover.Thisresearch explores the principles of conjoined lander‐rovers. The great benefit isminimizingcomponentcounts, todomorewith less. Thesepayoffsequate to lessmass, simplified integration, and lower costs than achievable with conventionalconfigurations. This enables capable planetary exploration from small launchvehicles.

Many configurations for capable surface exploration succeed with hundreds ofkilogramsofdrymass.Onlyconjoined lander‐roversmightcross theonehundredkilogram barrier. Results include the development of morphology, mechanisms,structure, and avionics to enable lightweight exploration. Developments detailspecializeddesigntohandle thehardvacuum, thermalswings, radiation,anddustonthemoonwithintherestrictionsoflowmassandvolume.

AnexemplarymissionusedascontextforthisisresearchistheGoogleLunarXPrize(GLXP).TheGLXPoffersa$20milliondollarprizetosendarobottothemoonandtransmitbackhighqualityvideoandimagery.

AcknowledgementsThisresearchharnessestheworkoffaculty,staff,andstudentsatCarnegieMellonUniversityinpursuitoftheGoogleLunarXPrize.Withouttheirtirelesscontributionsthisbreadthofresultswouldnotbepossible.In addition, Astrobotic Technology, Raytheon, University of Arizona, LockheedMartin,andNASAprovidedassistance.ProfessorRedWhittaker,theadvisorofthisproject,deservesspecialthanksforhisleadershipoftheXPrizeandhisconstantdedicationtoresults.

3

TableofContents1. Introduction ...................................................................................................................... 51.1. Motivation ............................................................................................................................... 51.2. Challenge..................................................................................................................................6

2. Background ....................................................................................................................... 72.1. Assumptions ........................................................................................................................... 72.2. MissionMass........................................................................................................................... 92.3. BlastoffMission .................................................................................................................. 102.4. Lunokhod.............................................................................................................................. 112.4.1. TheMorphologyofLunokhod ................................................................................................112.4.2. TheMechanismsofLunokhod................................................................................................122.4.3. TheStructureofLunokhod ......................................................................................................132.4.4. TheAvionicsofLunokhod ........................................................................................................13

2.5. Pathfinder/Sojourner....................................................................................................... 132.5.1. TheMorphologyofPathfinder ...............................................................................................142.5.2. TheMechanismsofPathfinder...............................................................................................152.5.3. TheStructureofPathfinder .....................................................................................................152.5.4. TheAvionicsofPathfinder .......................................................................................................15

3. ThesisStatement ...........................................................................................................16

4. Morphology .....................................................................................................................174.1. IntegratingLanderandRoverFunctions ................................................................... 174.1.1. AdvantagesofConjoinedSolution ........................................................................................18

4.2. PropulsionMethodology ................................................................................................. 184.3. SolarandThermalConfiguration ................................................................................. 194.3.1. ComponentPlacement ...............................................................................................................21

4.4. MissionControl................................................................................................................... 224.5. CameraConfiguration....................................................................................................... 244.6. CommunicationMethodology........................................................................................ 26

5. Mechanisms.....................................................................................................................285.1. LandingMechanisms ........................................................................................................ 285.2. MobilityMechanisms........................................................................................................ 295.2.1. DriveActuation .............................................................................................................................315.2.2. DriveSuspension..........................................................................................................................335.2.3. Traction ............................................................................................................................................34

5.3. CameraMechanisms ......................................................................................................... 345.4. RegolithProtection ........................................................................................................... 35

6. Structure...........................................................................................................................376.1. StructuralLoading ............................................................................................................. 376.2. CompositeMaterials ......................................................................................................... 386.2.1. RoverChassis .................................................................................................................................396.2.2. LanderStructure...........................................................................................................................406.2.3. SecondaryStructures..................................................................................................................41

4

6.2.4. CompositeThermalRegulation..............................................................................................427. Avionics.............................................................................................................................437.1. CommercialOff­The­Self(COTS)Avionics ................................................................. 447.2. PoseSensing ........................................................................................................................ 457.3. MotorControl ...................................................................................................................... 467.4. Software ................................................................................................................................ 46

8. Conclusion........................................................................................................................488.1. Claims..................................................................................................................................... 488.1.1. Configuration..................................................................................................................................488.1.2. Morphology.....................................................................................................................................488.1.3. Mechanisms ....................................................................................................................................498.1.4. Structure...........................................................................................................................................498.1.5. Avionics ............................................................................................................................................49

8.2. FutureWork......................................................................................................................... 509. PerspectiveandContribution ...................................................................................52

10. References.....................................................................................................................53

5

1. Introduction1.1. MotivationThisinvestigationdevisescapablelunarlander‐rover exploration fromminimallaunchcapability.Thegreatchallengeofexploration robotics is configuringmorphology,mechanism,structure,andavionics relative to the constraintsimposed by missions. The largestconstraint in the design of planetarylanders and rovers is typically themethod of getting from Earth to thedestination. Big rockets enable landersand rovers with large mass anddimension,butcomeathighcost.Smalllaunch vehicles reduce costs, butpresent the need for ingenuity toproducedesignsof lowmass and smalldimension that are capable of meetingmissionrequirements.

Withtheresourcesofnations,mightyrocketsoncecarriedimmensehardwareintospace. Ambitions for lunar exploration are re‐invigorated but huge launch andinjection stages are unaffordable,which compels success onmuch smaller launchvehicles. The burning question for robotic lunar missions is whether substantialexploration capability can be delivered with a class of lander‐rovers with dry,landedmassoflessthan120kg.Decades have passed, technical evolution has refined landers and rovers, andimprovement has been continuous, but ongoing evolution offers no leap ofperformanceforagivenlaunch.Aslongaslandersandroversaredistinct,thereisinevitableredundancyofstructure,power,computing,andcommunication.Aslongas roversarepayload, theyarenecessarily compromised inexploration capabilitybyinordinateconstraintsonmassandscale.Fortraditionallandersandrovers,eachredundant component is manifested in mass and volume, which precludesminimizationofamobileplanetaryexplorationsystem.Thisresearchexplorestheprinciples of conjoined lander‐rovers. The opportunity is to eliminate boundariesdemarcatinglanderandrover.Thegreatbenefitisminimizingcomponentcounts,todo more with less. Payoffs equate to less mass, simplified integration and lowercoststhanachievablewithconventionalconfigurations.

Figure1:ConjoinedLander­Rover

6

1.2. Challenge

Thisresearchdetailsthedevelopmentofconfigurationforconjoinedlander‐roversto explore the lunar equator. These exploration systems are intended for short‐durationscoutingmissionstothelunarsurfaceasprecursorstohumancolonizationofthemoon.Short‐durationisdefinedtobeonelunardayorfourteenEarthdays.

Thelunarequatorpresentssurfacetemperaturesfromaslowas‐170°Cduringthelunarnighttoashighas120°Catlunarnoon.Overthecourseofthelunarday,thesun traverses fromhorizon at dawn to directly overhead atmidday to horizon atdusk. Mission length of a single lunar day removes the requirement of survivingduring the fourteen day, ‐170°C lunar nights, which current space electronicsystems cannot survive. This drives requirements to operate continuously and athigherspeedsthantraditionalspaceroboticsystemstoperformallrequiredactionsinlimitedmissiontimes.AnexemplarymissionusedascontextforthisresearchistheGoogleLunarXPrize.TheXPrizeoffers$20millionforanon‐governmentalgrouptosendarobottothemoon and transmit back high qualityvideo and imagery. To compete forthisprize, the scaleof assets availabledrives the need for a small launchvehicle.Thisdictateslimitationsonthemass and dimension of a lunar landerand rover. The mission concept is towin the XPrize by visiting the site ofthe Apollo 11 landing at the lunarequator. This is seen as the bestlocation to inspire the public andincrease excitement of lunarexploration.

Figure2:Apollo11site

7

2. Background2.1. AssumptionsThe goal of this research is to achieve capable exploration from a small launchvehicletocompetefortheGoogleLunarXPrize.Thisrequiresdefining“capable”anddeterminingareasonabletargetfor“low‐mass.”The lander‐rover for the exemplary mission to compete for the XPrize must becapableof:

• Landingwithin100metersofapredeterminedsite,

• Trekkingtenkilometeronthelunarsurfaceinonelunarday,

• Collectinghigh‐definitionstereovideoduringtrek,• Communicatingtoreturndatacollectedoverthecourseofthemission.

Themassofanexplorationsystemderivesfrommissionrequirementsandpayloadrequired for success. Mission requirements dictate data to be collected oroperations to be performed. Payload takes the form of scientific instruments,imagingsystems,ortoolstoperformoperationslikesitepreparation.TheXPrizemissionplaces requirementson imagery tobe retrievedandreturned.Commercial missions add requirement for additional mass to carry contractedpayload as a source of funding. Once mission requirements and payload aredetermined,landerandroversystemsaresizedintheformofamassbudget.

The largestdriver indeterminingmass required formission success is thepowerrequired. Data processing, sensing, communication, and mechanisms are sizedbasedonmissionrequirements.Eachofthesedrawsasetamountofpowerrelativeto their performance and duty‐cycle. As power usage escalates, the size of solarpanels and radiators escalates and the lander‐rover requiresmore structure andmoremobility.

From mission requirements, appropriate electronics are selected and roughcalculationsforsystempoweraremade.TheresultfortheXPrizeisasystemwithcontinuous operating power of 120 watts. From operating power and duty cyclesolar and thermal control surfaces are sized setting the scale of structure andmobility from which an approximate mass is determined for the lander‐rover. Atargetmassbudgetof120kgwasdeterminedbasedoffthisinitialsizing.Asamplemassbudgetfora120kg,120‐wattequatorialconjoinedlander‐roverfortheXPrizemissionisshowninTable1.

8

Lander­RoverSystem

Function MassBudget(kg)

Payload Fulfillmissionrequirements 9

Avionics Processandstoredata 7

Communication Transmitdata 4

SensingLocalizelander

Localizerover6

MechanismPropelroverPointandplaceinstruments

15

Power Generate,store,anddistributepower 8

ThermalControlRadiateexcessheat

Heatcoldcomponents6

Harness Distributedataandpower 8

Structure Resistlaunch,transit,landing,anddrivingloads 35

Propulsion SlowdownforlandingPositionforlanding 22

TotalSystemMass: 120

Table1:MassBudgetfor120kgconjoinedlander­roverexplorationsystem

Success can be achieved with lower mass components, though this increasesmission risks. For example, lowermass Earth electronics can be used in place ofheavy space‐proven designs at the cost of increased risk of mission failure.Thereforeanappropriatebalancemustbestruckbetweenriskandreward.FortheXPrizemission customdevelopments areminimized for electronic components toreduce schedule and cost impacts. However, utilizing existing, newly developedhardwarethathasbeentestedonEarth,butwithlittletonospaceflightheritageisconsidered appropriate to achieve mass savings. Mechanisms and structure arecustomdevelopments, butharnessmaterials, lubrications, actuators, andbearingsthathavebeenproveninaspaceenvironment.

9

2.2. MissionMass

Mission hardware must launch intoEarth orbit, propel itself toward themoon, brake to match the moon’sspeed upon arrival, and navigate toland on the surface. Every kilogramonthelunarsurfaceback‐propagatesintokilogramsofpropellantrequiredtolaunch,inject,brake,andland.Theresult is large increases in launchvehicle requirement and in overallcostformissionsuccess.EachstageofthetripfromtheEarthtothemoonisaccomplishedthrougha change in velocity. Reaching Earthorbit requires accelerating to acertain velocity. To reach the moonfromEarthorbit, spacecrafts changevelocitywhenaccelerating toward themoon,brakinguponapproachofthemoon,andslowingdownandlandingatanintendedlandingsite.Requiredchangesinvelocitydependonorbitalmaneuversperformed.ValuesforasamplelunarmissionareshowninTable2.

Using these delta velocities, the mass of propellant required can be determinedusingtheTsiolkovskyrocketequation[13]:

€

Δv = ve lnmfuel + mdry

mdry

Inthisequation

€

ve istheeffectivevelocityof fuel exhaust,which is ameasureof theefficiency of a type of fuel to providechanges in velocity relative to expendedfuelmass. There are twomain categoriesof rocket fuel: solid and liquid. Solid fuelshave less complexity and associated tankmass than liquid fuels resulting in higherpropellant mass relative to tank mass,which is exaggerated in smaller rockets.Solid rockets lack the controllability ofliquid rockets. Once ignited, solid rocketsburn the entirety of their fuel at once,while liquidrocketscanbe turnedonand

MissionStage Delta­velocityrequired(meters/second)

LaunchtoLowEarthOrbit1 10,000

Injection2 3100

Braking2 2200

Landing 400

1:BasedonLunarProspectormission[14]2:Calculatedusingorbitalmechanics[15]Table2:Requiredchangesinvelocityduring

missionphases.

Figure3:Missionstackinsidealaunchfairing

10

off for controlled burns. Launch, injection, and braking workwell with solid fuelsincetheyrequirelargechangesinvelocitywithnoneedforspecificcontrol.Orbitaladjustmentsandlandingworkbestwithliquidfuelsincetheyrequiresmallchangesinvelocityandneedspecificcontrol.

FullmissionstackupsaregovernedbytheTsiolkovskyrocketequation.Propellantrequirementsforeachstagecanbedeterminedbystartingwithdry,landedmassonthemoonandworkingbacktoEarth,addingonstructuralmassforeachstagealongthe way. Table 3 steps through this process for a 120kg system used for theexemplary XPrize mission discussed in this research. In this example the launchvehicle is a four stage Athena II rocket that was used for the Lunar Prospectormission,a1998lunarorbiterwithsimilarmasssenttothemoon[14].

MissionComponent RequiredMass(kg)

MassonMoon 120

Propellantforlanding1 18

BrakingStructure2 40

PropellantforBraking1 208

InjectionStructure2 130

PropellantforInjection1 1,018

LaunchStructure3 9,800

Propellantforlaunch3 108,000

TotalLaunchedMass 119,3341:DeterminedusingtheTsiolkovskyRocketEquation2:BasedonmassofATK’sStarmotorline[4]3:NumbersfromtheAthenaIIlaunchvehicle[4]Table3:Massbudgetfor120kglander­rovermissionto

themoonbasedondeltavelocitiesinTable2

2.3. BlastoffMissionTheBlastoffCorporation formulatedamobile landedmission froma small launchvehicle similar to thatdiscussed in this research.Blastoffpursueda commerciallyfundedmissiontotheApollo11landingsite.

TheBlastoffmission configuration did not reach a technical solutionwithinmassconstraintsoftheintendedAthenaIIlaunchvehicle.Achangewasmadetoalargerlaunch vehicle at higher mission cost. Eventually escalating mission costs and

11

collapse in financial backing resulted in abandoning the mission. Successfulcommercial lander‐rover exploration from small launch assets requires theadvancementofconfigurationforlowmass.

Figure4:Blastofflander­roverconceptforexplorationofthelunarequator[31]

2.4. LunokhodThe first planetary lander and roverweredevelopedbytheSovietsin1970forexplorationofthemoon.Principlesfrom that era still governdesign. Theidea of distinct and separate landerand rover originates from the SovietUnion’s Lunokhod campaign. TheProton rockets that launched theLunokhods were capable of launching5,600kgdirectlytowardthemoon.TheLunokhodswere800kgroverson800kglanderswithcombineddimensionsof3.3metersindiameterand2.3metersinheight.The Lunokhod rovers and Luna landers were distinct and separate systemswithdistinctmechanisms,structure,andavionics.Theystackeduplikeaweddingcakeinthe Proton fairing and separatedmechanically once on the surface on themoon.This presented the need for many duplicated systems: two computers, twocommunicationsystems,twolocalizationmethods,andtwoindependentstructures.

2.4.1. TheMorphologyofLunokhod

TheLunokhodroverchassiswasamagnesiumshell,akintoacoveredbathtubthatwas overlaid with a large hinged, convex solar panel lid. The rover trekked thesurface on an undercarriage of eight large, widespread wheels giving theappearanceofamilitarymachine.Cameras,antennas,andsensorsprotrudedfromthetubstructure.Thisappearanceservedasdistinctandinnovativesolutionstotheproblemsofthelunarenvironment.

Figure5:LunokhodroveronLunalander

12

The bathtub chassis was a pressurevessel sealing in an atmosphericenvironment of nitrogen for theavionics of Lunokhod. This shieldedcomponentsfromtheharshvacuumofthe moon facilitating thermalregulation. The hinged solar panelopenedduringthelunarday,gatheredpower from the sunand revealed theradiative top surface of the chassisenabling the rover’s avionics to cool.The lid closed during the long coldlunar night covering the radiativesurface to insulate the rover. During

the night, heat generated from the decay of a radioactive isotope kept the roveravionicswarm.

Theundercarriagegavemobilitytotheplanetaryrover.Thewidestanceandeightwheels improved the obstacle and slope performance of the rover, producing amachine ready to handle almost any situation that may occur. The use of eightwheelsensuredlocomotioninthecaseoffailureofasinglewheel.

2.4.2. TheMechanismsofLunokhod

The Lunokhod design incorporated mechanisms of eight independently drivenwheels,torsion‐barsuspension,solarlidtilt,andantennapointing.Theundercarriagehadeight rigiddrive‐wheelswithperforated, cleated rims.Thewheelsdidnotswivel.Tank‐likesteeringimparteddifferentvelocitiestotheleftandrightsides,turningtherover.Brushed,directcurrentmotorswithplanetarygearinglocatedwithineachwheelhubactuatedlocomotion.Wheelhubswereindividuallypressurized and sealed with custom vacuum grease, creating an atmospherenecessary topreventsparkingatbrushes thatcauseswearandeventual failureofbrushedmotors inavacuumenvironment. A shifting transmission in thegearingenabledtwospeedsof1km/hrand2km/hr. Intheeventofactuationfailure inawheel, an explosivedisconnect couldbefiredtoenablethewheeltofreespin[1].

Similaractuators facilitatedthepointingof Lunokhod’s high gain antenna. Thissystem was manually steered fromEarth.Thesolarlidwasactuatedtopointtowardthesun.

Mechanismswerealso inplace to covercameralensestoprotectthemfromdustkicked up during landing. These coverssimply dropped away once on thesurfaceofthemoon.

Figure7:Lunokhoddrivesystem

Figure6:LunokhodRover

13

2.4.3. TheStructureofLunokhodSpeciallightweight,highstrengthmagnesiumandtitaniumalloysweredevelopedatRussia’s Lavochkin Institute for use in Lunokhod [8]. The development ofspecializedmaterials to reducemass highlights the importance ofmass reductioneveninasystemwithacomparativelylargemassbudget.

Thestructuresoftheroverandlanderwerecompletelyseparateandsimplystackedon top of each other. Rover structure consisted of a pressure vessel chassiswithtitanium strut mounts supporting an array of cameras, antennas, and sensors.Pressurizedcompartmentssucceededwiththinmagnesiumcasings.

2.4.4. TheAvionicsofLunokhodThepressurizedcompartmentofLunokhodprotectedavioniccomponentsfromtheharshvacuumonthelunarsurface,enablingcomponentstobethermallymanagedby forced air. Availability of large mass enabled components like cameras to beshieldedfromradiationeffectsenablinguseofcommercialhardware.

Lunokhodwasequippedwith fourTVcameras fordrivingandreturning imagery.Threewerepanoramiccameras,andthefourthwasmountedhighontheroverfornavigation.ThenavigationcamerawasaddedtoLunokhod2duetothedifficultyofoperators to observe terrain from low mounted panoramic cameras resulting indrivingintoanunseencrater,causingthefailureofthefirstLunokhodmission[7].

Afive‐manteamofcontrollersdrovetheroverinrealtime.Communicationstoandfromtheroverweresentthroughalowdataratecone‐shapedomni‐antennaandahighdataratedirectionalhelicalantenna[3].Drive motors had automatic motion‐control electronics.Rotationalsensorsprovidedfeedbackforwheelmotion.Afree‐rolling wheel provided information on the distancetraveled and determined wheel slip when compared tocountsofwheelrevolutions[2].

Scientific instruments included a cone penetrometer totest soil mechanics, a solar X‐ray experiment, anastrophotometertomeasurevisibleandUVlight levels,amagnetometer,aradiometer,andaphotodetectorforlaserdetection experiments. It also carried a laser corner‐reflectortoenableprecisemeasurementoflocationonthemoon and the exact distance between the Earth and themoon.

2.5. Pathfinder/SojournerThe Mars Pathfinder mission was the first mission to investigate strategies forcombining the functions of lander and rover. Hardware included the PathfinderlanderandtheSojournermicro‐rover.Successoftheprogramwasachievedthroughlowmassand lowcostrelative topriorNASAmissions.Thesystemhadamassof

Figure8:Afree­rollingwheelwasusedtodeterminewheelslip

14

275kg and cost $280 million [10]. Themajorityof systemswere locatedon thelander with minimal capabilities on therover for short distance traverses andcollectionofscientificdata.

The Martian environment is less harshforexplorationrobothardware than themoon. The presence of an atmospherelimits thermal swings, enablesconvective cooling of components, andcreatesdiffuselightingandsolarenergy.The short Martian night mitigatesextreme low temperatures. Terrain islessextremeandcolorationand lighting

facilitateeasieridentificationofobstacles.The Pathfinder mission succeeded with low mass due to limited explorationrequirements. The Sojourner rover never trekked more than 10 meters fromPathfinder.Landerandrovershared tasks toreduceduplicationofhardware.Themethodologyusedplacedthemajorityofcapabilityonboardthelanderandlimitedtherovertooperationswithinline‐of‐sightwiththelander.

2.5.1. TheMorphologyofPathfinder

The Pathfinder lander came into Mars as a pyramid and opened like a flowerrevealingthreesolarpetalsandacentralavionichousingandcommunicationcenterwithavarietyofscientificinstruments.TheSojournerroverwasperchedononeofthe solar petals and looked like a boxy remote control car. A flat, upward‐facingsolarpanelmadeupthetopsurfacewithagoldchassisandsixwheeldrivesystemsuspendedbythinaluminummembers.Limitedcapabilitiespermittedasmallroverofonly10.5kg.

The Pathfinder lander slowed on approach toMarsthrough a heat shield to brake against the Martianatmosphere and through parachutes. Impact wasabsorbedbyairbags.Heatshieldsandparachutesdonotworkonthemoonduetothelackofatmosphere.These systemsalso led toan imprecise landing site,notsuitedtoamissionthatseekstolandwithinafewkilometersoftheApollo11site.

Diffuse lighting, minimal operation time, and anequatorial location enabled the use of the upward‐facing,non‐actuatedsolarpanelsonboththe landerand rover. A charge‐operate cycle permitted therover to draw more power in operation than itproducedfromsolarcells.Theroveronlyoperatedamaximumoffourhoursaday.Avioniccompartments

Figure9:PathfinderLanderwithSojournermicrorover

Figure10:Sojournerrover

15

have minimal heat rejection requirements in the relatively moderate thermalenvironmentofMars.The Sojourner rover served as short‐range data collection, transmitting andreceiving data through the lander, which communicated with Earth. Six‐wheeledlocomotion with a rocker‐bogie suspension provided excellent rough terrainmobilitytotheslowmovingrover.

2.5.2. TheMechanismsofPathfinder

Six wheels propelled sojourner with brushed motors and planetary gearingcontainedwithineachwheel.Thetwofrontandtwobackwheelswererotatedusingfour individualactuatorscausingtherovertoturn.Theatmosphereandmoderatetemperatures ofMars facilitate the use of less advanced grease and oil lubricatesandbrushedmotors,whicharenotpossibleinthehardlunarvacuum.Roverspeedtoppedoutat.036km/hr[11].Sojournerusedapassivemethodofdifferencingknownasrocker‐bogie.Suspensionjoints rotate and conform to the contour of the ground, providing stability fortraversingrocky,unevensurfaces.Thissystemenabledtherovertoclimbrocksuptotwicethediameterofitswheels.

2.5.3. TheStructureofPathfinderSojourner‘schassiswasprimarilyaluminum.Actuatorhousingsandbearingswereintegrated into the chassis to reduce mass for appended components. Lightweightedaluminumlinkageswereusedintheroversuspensionsystem.

2.5.4. TheAvionicsofPathfinderAvionics in both the lander and rover were packaged tightly together andsurroundedbyalayerofAerogel,anextremelylightinsulatingmaterial,tomaintaintemperature of components for minimal mass. This strategy worked well in thecolderenvironmentofMarswhereheatingistheprimaryconcern.

The Pathfinder lander included communication to Earth, an array of sensors,computing,andcameras.Thelanderandrovercommunicatedthroughmodemsforline‐of‐sight communication. To save on cost themodems usedwere commercialhardwarewithminimalin‐houseadaptationsforspaceflight[12].Sojourner included minimal avionic components as most computing was locatedonboard the lander. These included a light‐striping system for navigation. Thissystemusedfixedlaserrangesensorstoscantheterrainimmediatelyinfrontoftheroverforautomateddetectionofobstacles.TherovercarriedanAlphaProtonX‐raySpectrometer to analyze the composition of rocks. It was also equipped with acameratoaccessthelander.

16

3. ThesisStatementLunokhod and Sojourner solutions do not provide low‐mass capable exploration.Lunokhod succeeded with high mass and high mission costs. Sojourner lackedexplorationcapability.Bothsucceededwithlargelaunchvehiclesrelativetoa120kglander‐rover.

This research devises capable lander‐rover systems that meet mission objectiveswithin the restrictions of extremely small launch vehicles. The configuration of a120kg, conjoined lander‐rover is investigated in the framework of a mission tocompete for the Google Lunar XPrize. This is accomplished through developinginnovativemorphology,mechanism,structure,andavionicstominimizedimensionandmass.

17

4. MorphologyThemorphology of a space exploration system is the form it takes based on theconfiguration and placement of components. Morphology for a conjoined lander‐roverincludesthemethodologiesforintegratinglanderandroverfunctions,andtheconfiguration of propulsion, power, thermal, control, camera, and communicationsystems.

4.1. IntegratingLanderandRoverFunctionsDevelopingconfigurationsforconjoinedlander‐roversrequiresdeterminingwhereto draw the lines demarcating lander and rover. An extreme is a rover withthrusters appended to the chassis and mobility systems that double as landingappendages.Thiseliminatesduplicationofcomponents,as the lander is therover.Roving capability, however, is hampered as propellant tanks and propulsion addbulk and mass to the mobile system. The lack of a separate lander means themobility system and rover underbelly are not shielded from dust kicked up onlanding.Aconfigurationwithremarkableadvantagesreducesthelandertoaplatformwithpropulsioncomponentsarrayedaboutitsbase.Propulsioniscontrolledbyavionicson the rover through an electrical umbilical connection.All avionics and softwarereside in the rover, and the rover flies the propulsive‐only lander to the surface.Onceon the surface, the rover separates from the landingpallet anddrivesaway.The true distinction is that the lander is bereft of avionics, and that the roveravionicsaretheonlyavionics.Collateraladvantagesarecapabilitiesfordual‐useofstructure,avionics,perception,communication,andcontroltoachievecapable,ultralightsurfaceexploration.

Figure11:Devisedconjoinedlander­roverconfiguration;avioniccomponentsresideintherover;thelanderisonlyapalletwithpropulsionappended.

18

4.1.1. AdvantagesofConjoinedSolutionConjoining the systems of landers and rovers facilitatesmeetingmass budgets incapable exploration systems. Conjoined lander‐rovers have significant advantagesover traditional split architectures in the areas of structure, avionics, perception,communication,andsoftware.

Structure: Stack height, fairing diameter, and taper are all constrained in smalllaunch vehicles. A discrete, wedding‐cake stack of lander and rover cannotminimize vertical height. The advantages of monolithic design reduce stackdimensions and achieve an overall mass advantage by reduction of elementalsectionsandreductionofmaterialvolume.

Avionics: Traditional landers and rovers are distinct spacecraft that requireseparate avionics hardware for control, power and thermal management, andcommunications.Combininglanderandroverfunctionsenablesthesystemtomakedual‐useofonecomputer,onepowersystem,andonecommunicationssystem.Thissingleavionicssystemsimplifiestesting,reducesthecableharnessingrequired,andsavespowerandmassthatisduplicatedinsplitarchitectures.Perception: Cameras and an Inertia Measurement Unit (IMU) are mainstays ofsensing for rover exploration. Radar and an IMU are the mainstays for entry,descent, and landing (EDL) range and pose estimation. The system envisionedmakes use of one IMU, activewhile landing andwhile roving. Cameras guide forroving, but are also used to match imagery captured during landing to satelliteimagerytodetermineposeandnavigateforapinpointlanding.Communication:Communication for landing and roving are traditionally distinctsystems. The innovation in a conjoined system is a monolithic communicationsystemthathasonlyonetransponderandmultipleantennasforlandingandroving.

Software:Software,firmware,andcomputingforflighttraditionallyareintrinsictoflight, and inactive once a spacecraft reaches the surface. Software, firmware, andcomputingforrovingtraditionallyareintrinsictoarover,andinactiveduringflight.Necessaryutilitysoftware likedevicedrivers,operatingsystems,communications,powermanagementand thermalmanagementarecommon ina split lander roversetup.Thiscommonalityisexploitedinaconjoinedlander‐rover.Byreducingtheamount of redundant code, the spacecraft flies with fewer lines of code that aredevelopedinashorterperiodoftimeandtestedmorerigorously.

4.2. PropulsionMethodology

Largemass systemswith an excess of propellant succeedwith long rocket burnsimpartinglowaccelerationloadsonspacecrafts.Smallmassmarginsleadtorocketburns with short durations and high accelerations to improve efficiency ofpropellantuse.

Lightweight missions exploit the short forceful burn methodology, utilizingpowerful solid stages to impartmostof thenecessary changes invelocity inburn

19

timesofonlyoneortwominutes. Liquidbrakingthrustersonthelanderperformthe final slow down, bringing the craft to near‐zero velocity relative to themoonwithafewstrongshortbursts.

Small attitude control thrusters on the lander adjust direction to hone in for apinpoint landing. These thrusters slow the craft to a slow constant velocity a fewmetersbeforetouchdownonthelunarsurface.

Innovative low mass options such as ion propulsion and solar sails provide thepossibility for future systems to traverse from Earth orbit to their intendeddestinations in low mass packages. The downside is these processes work byincrediblyslowchangesinvelocityexpandingafourdaylunarjourneyintoseveralmonthsormore.Thisaddscosttothemissionintermsofincreasedlifetimeofpartsandincreasedtimelineforuseofcommunicationcenters.

4.3. SolarandThermalConfigurationLander‐rovers for lunar exploration must provide power and thermal control inflight,aswellasonthesurface.Solarpowerisreadilyavailableduringthelunardayandisaninexpensive,safe,andlowmasssolutiontoprovidecontinualpoweroverthe course of a mission. Solar panel design for rovers traversing the moon isdifferentfromthatofsatellites.Aconstantlychangingsolarincidentangleduetotherover’smovementandsunelevationcontinuallychangesthepowerprofile. Atthelunarequator,thesuntreksacrosstheskyfromhorizontooverheadtohorizonandthelackofanatmosphereeliminatesalldiffusionofsolarenergy.Thisrequiressolarconfigurations to capture solar energy originating from a variety of angles forcontinuous operation. Similarly, radiative surfacesmust account for the changinglocationofthehotsun.Lunokhod solved solar and thermal challenges using an actuated solar panel topointtothesunatanyangleandaradiatorfacingdirectlyup.Thissolutiondoesnotsolvetheconfigurationforlightweightexploration.Actuationaddsmechanismandmass.Upwardfacingradiatorsprecludeactionatmiddaywhenthesunisoverhead.

A symmetric configuration for equatorial rovers seeks to provide power andthermal regulation with any direction of travel with no actuation. The form is asymmetricfour‐sidedpyramidwithsurfacesinclinedatthreedifferentangles.Solarpanelsinclinedat90°fromthegroundcollectpowerintheearlyandlateday,whilesteeplyinclinedsolarpanelscollectpowernearmidday.Radiatorsurfacesareatanintermediate angle to reject heat evenly regardless of which side the sun is on.Refinementplacedallsolarcellsatoneinclination,removing90°panelsthatpulledintoomuchheatfromthesurface.Symmetricconfigurationsareinfeasible.Radiatorsurfacesonthesunsideaggregatemoreheatthanothersurfacescancompensate.Componentsmountednearthesunside would not maintain thermal limits. Solar cell mass would be significant toachieve enough power when driving in any direction, since at least half of cellswouldgounusedinanyorientation.

20

Asymmetric configuration succeeds relativeto solar and thermal challenges. Whenmobile on the surface, rovers drive withsolar panels into the sun for power, withradiatorpointingawayfromthesun,towardblack space for cooling. The inclined facetson one side gather solar power during thechangingsunanglesthatoccurfromdawntodusk.Anupward‐lookingfacetontheothersideejectsheattoblackspace. Thebalanceofthesolarheatingandradiativecoolingaremeans for the essential regulation of thetemperaturesofinternalelectronics.Duringa rolling cruise, the same arrays and

radiatorsgeneratepowerandregulatetemperatures.Theinnovationistogeneratepower and regulate temperature both in cruise and on the surface withoutduplicationofcomponentsoractuationofsolararraysandradiatorsurfaces.

Thesolararraysregulate theirowntemperature.This isdonebyconstructing thestructure of the four solar facets out of a continuous piece of highly conductivecarboncompositetodistributeheatawayfromilluminatedpanels.Cellsthatarenotilluminatedactasradiators,rejectingheattoblackspace.

Configurationofpowersystems for conjoined lander‐rovers includesoptimizationof panel angles, cell string design, and charging system. Four facets permitorientationatarangeofanglesrelativetothesunwhilestillgatheringpower.Theabilitytotravelineitherdirectionisachievedbyreversingwheelspin.Resultsofasimulationquantifyingpoweratdifferent sunelevationanglesanddifferent roverdriving angles relative to the sun for a sample solar configuration developed areshowninFigure14.

Figure13:Evolutionofsolarandthermalconfigurationfromasymmetricsystemtoalowermass,optimizeddirectionalrover

Figure12:Analysisofthermalconditionsforasymmetricconfiguration.

21

Figure14:Poweratdifferentsunelevationanglesanddifferentroverheadingsrelativetothesun.Aheadingof180° indicatesthesolarpanelsarepointedatthesun.A0° elevationiswhen

thesunisonthehorizon.

4.3.1. ComponentPlacement

Placement of avionic components is determined largely by thermal requirements.The connection between components and radiator surfaces determine thetemperature range experienced. The more direct the contention, the closer thecomponenttemperatureistotheradiatortemperature,andthemoresensitiveitistofluctuationsinradiatortemperature.Density of component packing impacts thermal performance. Close packing helpscomponents share heat in a cold environment, while spreading out componentsfacilitatesradiativeheatrejection.Equatorialroversexperiencetheextremelyhightemperaturesofthelunarday.Thisresultsinaconfigurationwithcomponentsconnectedeitherdirectlytotheradiatoror through high conductivity pathways. Components are distributed across theradiatortoimproveefficiencyofheatrejection.

22

4.4. MissionControlLunar rovers are tele‐operated from Earth through a time delay and requireconstant attention from their rover operators. In order to achieve the missionobjectives in a timely and cost effective manner, rover operations must be fast,flexible, and efficient. Because of the expense of communication bandwidth forremote operations, small improvements to the efficiency and safety of rovers canleadtolargemissioncostsavings.

Terrestrial mission control designs traditionally form tight loops between robotsand human operators [24]. They require many people to analyze an inordinatevolumeofdata returned from thevehicle [25].Time toanalyzedata impingesondecision‐making. Successful configurations of mission control for lunar roversoptimizes presentation of data to operators and assists in the decision makingprocess.Rovercontrolleveragesexistingtechniquestointegrateuserinterfacetools,aidingin decision‐making. This includes systems to visually suggest next actions, slidingscale autonomy, and to providethree‐dimensional models. What islacking in currentmission control isseamless integration of these toolsand techniques into operator workflow[27].Theinnovationofrovercontrolistheimprovementofremoteoperationbychangingtheway inwhichtheroveroperator interacts with the rover.The control system developedpresents rover interfaces that aresimple, effective, and safe such that,withminimal training,operatorscan

Figure15:Avioniccomponentsareconnectedtotheradiatoreitherdirectlyorthroughhighconductivitypathways.Distributionacrosstheradiatorimprovesefficiencyofheatrejection.

Figure16:Terrestrialmissioncontrolcenterspresentinordinatevolumesofdatatooperators.

23

directly control the rover. Themajoradvancementisthedesignand integration of tools forpredicting vehicle safety inresponsetocommandsissuedbyoperators.

Testing on terrestrial roverprototypes is essential to therefinement of remote controlcenters. These improvementscome in the form of reducedvolume of data presented tooperators by the formation of integrated tools that present meaningful dataproductstooperatorsbaseduponwhatismostcommonlyused.

Predictive views of trajectories improve tele‐operation of rovers. Operators areprovidedwithestimatesoffutureroverpositionsoverlaidonstereoimagery.Theseestimatesarebasedon inertialnavigationsystems,cameraangleandposition, theforeshortening of the image, and an estimate of the position of the terrain.Operators control waypoints and predictive modeling produces the intermediate

steps. If terrain is rough,operatorscanchoosetoslowdowninwhichcasethechainofstepsmoveclosertogetherandthe uncertainty lowers. This enablesoperatorstodynamicallychoosehowtomitigate risks given the structure oflocal terrain. Rover safety systems canalsooverrideoperatorcommands.Thisensures that, in the event of a suddencommunication loss or gross error inprojections,theroverwillbecapableofprotecting itself. This interactive riskmanagement increases systemefficiency and assists operators inassessing the advantages anddisadvantages of different navigationroutes.

Figure17:Flowofdatainmissioncontrolconfiguration

Figure18:Rovertrajectoryisoverlaidoncameraviewsbasedonpredictivemodeling.

24

4.5. CameraConfigurationMorphologymustbedesignedaroundtheprimarydeliverablesofthemission.TheGoogle Lunar XPrize is first and foremost about movie making. Desired imagerydetermines the type of camera systems and their placement. Cameras for thismissionmustbecapableof:

• Creatingpanoramicimagery,• Pointingineitherdirectiontoenabletwodirectionsofdriving,• Zoomingintoseehistoricalhardwarewithoutdisturbingthesite,• Viewingatleast40%ofthesurfaceoftherover,• Capturing3Dimageryandvideo.

Samplemorphologycombinesanupwardfacingcameraandabowlshapedmirrortocollectpanoramicimagery,asshowninFigure19.Forwardandbackwardfacingcamerasrecord3Dimageryandenabledrivingineitherdirection.

Figure19:Acombinationofanupwardfacingcameraandbowlshapedmirrorcollect

panoramicimagery;forwardandbackwardfacingcamerascollect3Dimagery.

Amodificationexchangesthebowlandupwardcameraforagimbaledcameraandstereo navigation cameras only on one side. The gimbaled camera collectspanoramicimagesandincludesazoommechanism,arequirementthatismissedinthe bowl‐mirror configuration. The gimbaled camera is also used to drivebackwards eliminating the need for rear navigation cameras. This succeedswhencombined with a symmetric solar and thermal configuration so driving in anydirectionispossiblebyfirstturninginplace.

25

Figure20:Aniterationofcameradesignintroducesagimbaledcameraandonlyplacesstereo

navigationcamerasononesideoftherover.

The optimal camera configuration for the XPrize is adapted to the asymmetric,directionalrover.Thisdesignplacesthreecamerasonapanandtiltsystemhighontherovermast.Onecamera includesazoommechanismandtheothertwocollect3D imagery. This enables any form of imagery to be captured on any side of therover.Thehighvantagepointisidealfordriverperception.

Self‐viewingisachievedbytheinclusionofsideviewmirrors.Thesemirrorspermitthe rover to see its drive system to assess danger and view a front logo panel toprovidetheinclusionofsponsorlogosimprovingmissionprofitability.

Figure21:Zoomandstereopaircamerasweremovedtopanandtiltmechanismatthetopof

therover.Mirrorsenabletherovertoseethelocomotionsystemandalogopanel.

26

Figure22:ViewfrommastcamerasastheXPrizeroverapproachestheApollo11site

4.6. CommunicationMethodologyThe morphology of communication systems for conjoined lander‐rovers mustprovidea linktoEarthduringallphasesofamission.Intransit,aspacecraftmustmaintain communication with Earth during all flight attitudes in order to sendtelemetry and receive commands for orbital adjustments. On the lunar surface,communicationmust dealwith changing rover orientation relative to Earthwhileprovidingdata rates capable of transmittingnecessarydata for operations and tofulfill mission requirements. Configuration of a communication system includesselectingantennason theexploration systemaswell asgroundstationsonEarth.The larger the antenna onEarth, the lower the requirement for antenna size andcommunicationpowerontheexplorationsystem.Communicationiscriticaltoanyroboticspacemission,butmoresofortheXPrize,sincethemissionrequiresvoluminousvideoandimagery.Thereisnoprecedentforachievingsuchcommunicationbandwidthfromaplanetaryrobot. Thisrequiresabreakthroughinlunarcommunication.

Asampleconfigurationusesalowrateomni‐directionaltotransmitdatarelevanttooperationsandasteerablehighgainantennatotransmitdatatoEarth.ThesystemconceivedwoulduseasmallcameratoidentifyandpointtotheEarth.

An optimal configuration for the XPrize takes advantage of cutting‐edge antennatechnology for surface operations. The technology selected is that of an evolvedantenna.Theseantennasappear similar tobentpaperclips, but the specificbendsare designed using machine learning techniques to optimize the antenna’s beampattern and efficiency for a given mission [15]. This beam pattern designencompasses the Earth regardless of rover orientation. This system continuallytransmitsdata foroperation interspersedwith imagery. Imagery isprioritized for

27

transmissionandtricklesinwhenexcessbandwidthisavailablethroughthecourseofthemission.

In addition to the roving antenna, two low power antennas reside on the landerdeck.TheantennasresideonoppositesidesandhavelargebeamwidthstoenablecommunicationwithEarthduringtransitregardlessofspacecraftorientation.

Figure23:Onedesignusesasteerableantennaguidedbyasmallcamera.TheXPrizeroversucceedswithalowmassstationaryevolvedantenna.

28

5. MechanismsMechanisms that succeed for lander-rovers have to fly, land, and rove subject to high loads within low mass and volume of small lunar missions. The harsh lunar environment necessitates unique mechanism design. Hard vacuum and extreme temperature ranges demand specialized lubrication methods. Thermalexpansionscauseinternalfrictionwithoutmeticulousdesign. Even the lunar soil (or regolith) possesses unique challenges to mechanisms, as it is made of small loosely packed, highly abrasive particles that pose significant contamination problems. Solutions must overcome these challenges within the constraints of limited power, mass, and volume.

5.1. LandingMechanismsLander mechanisms absorb landing impact and secure the rover during launch,cruise,andlanding.Theselowmassmechanismsmustbecapableofsurvivinghighloadsfromrocketsfiringandlandingimpact.

Landerlegsstowforflight,deployfordescent,andflexonimpact.Stowingfacilitatesmechanism survival during launch, cruise, and braking. However, successful legdesignmustendureaccelerationsfromlandingthrustersandlandingimpactwhilein a deployed state. Use of carbon composite materials in leg design improvesstrengthataminimalmass.Pressurizedpistonsabsorbimpactonlandingreducingloadstransmittedtothelanderandrover.

Legdesignandplacement facilitatestability.Three‐legdesignensuresall legswillbe incontactwith thegroundonuneventerrain,preventingrockingbetween legsduringroveregress.Widestanceinhibitstip‐overduringlanding.Duringflightandlanding,loadsaretransmittedbetweenlanderandroverthroughaseparationmechanism.Onceonthesurface,theroverseparates,thepalletlanderisinert,andtheroverdrivesaway.FortheXPrizelander‐roverseparationisachievedthroughtheuseofshapememoryalloys.Shapememory alloys arematerials thatproducelargeandpredictablechangesinsizeandshapewhenheatisapplied.Thelanderandroverareconnectedviaboltsnotched to fail at predeterminedlocations. Release mechanisms heatshapememory alloy materials to stresstheboltsintensionuntiltheybreak.Useof these mechanisms facilitates testingbyreducingdangerandincreasingreuseofcomponentsbetween testsrelative totraditionalexplosiveseparations.

Figure25:Shapememoryalloyactuatorsbreakboltsconnectingthelanderandrover[30].

29

Figure24:Onceonthesurface,theroverseparates,thelanderisinert,andtheroverdrives

away.

5.2. MobilityMechanismsOnceonthesurface,arovermustbecapableofmaneuveringoverterrainbetweenpointsofinterest.Anticipatedsurfaceterrain,speedoftravel,andlevelofsafetyinoperation are key factors in design. Amany‐wheeled extremely safe and capablesystemmeans addedmass anddimension. Solutions that succeededonLunokhodandSojourner areoverdesigned relative to the requirementsof ultra‐light rovers.Improvedabilitytosenseterrainandavoidmajorobstaclesfacilitatessuccesswithasimplifiedsystembycircumventinghazardousterrain.

Mobility mechanisms include actuation, suspension, and traction element design.Requirements of a drive system include vehicle speed, torque, power usage,traction,obstacleperformance,andslopeperformance.

In the design of mobility, sizing of gearing and motors ensures sufficientperformance.Asystemwithinadequatetorquecouldbecomestuckwithnomeanstorecover,strandingtherover.Significantsafetymarginsarerequiredtoprecludemechanism failure. The XPrize rover accounts for mission uncertainties throughdesignforfulloperationsonEarth.SincelunargravityisonesixththatofEarth,theroverwill have six times lessweight tomove around on themoon. Thismethodenables simplified testingwithoutgravityoffloading to testmissionoperationsonEarth.Thedownsideisadditionalactuationmassandlowerpowerefficiency,sinceanactuatorislesspowerefficientatlowertorqueoutput.Theprocessofsizingactuationbeginswithderivationofoutputpowerrequiredbydeterminingtorqueandspeedneeded(

€

P = Τω ).Powerrequirementsareincreasedby a factor representative of expected gearing efficiency loss.Motors are selectedthatsupplynecessarypower.Agearratiofortheactuatorcanthenbedeterminedbased on torque and speed ratios between motor specifications and desiredactuatoroutputs.Speedratioprovidesanupperlimitongearratioandtorqueratioprovidesthelowerlimit.Gearingisselectedtoachievearatiowithinthisrangethatiscapableofsurvivingtheanticipatedloadingcycles.Thisprocesscanbeiteratedasneededbasedontestedloadsandgearingefficiencyrelativetoassumptions.

Testing ensures mobility systems meet performance requirements. Obstaclesurmounting tests reveal obstacle performance. Power performance is measured

30

usingatestsetupto loadtheactuatorwithasettorqueusingamechanicalbrake.This measures actuator efficiency in terms of electrical power in relative tomechanical power out. Estimates of power are particularly important for spacesystemstoappropriatelysizepowersystems.

Figure26:Mechanicalbrakesetuptotestactuatorpowerefficiency

Traction performance can be assessed through testing on slopes or through adrawbarpulltest.Inthistesttherovermustdriveonflatgroundwithameasuredresistingforce.Thisiscomparabletotheresistancecauseddrivingupaslope.Thepercentageofwheelslipisplottedagainstresistiveforceasapercentageofvehicleweight. These results can be translated to traversable slope angles. Results areinterpolated to determine a range of safe slopes, dangerous slopes, andinsurmountableslopes.AnexampleresultisshowninFigure27.

Figure27:Graphresultingfromadrawbarpulltest.Slipratioisplottedvs.resistanceforceasa

percentageofvehicleweight.

31

5.2.1. DriveActuationActuationforadrivesystemmustprovidethepowernecessarytopropelandsteeraroveronthelunarsurface.Excesspowercapabilityisdesignedtohandleextremesituations. Slippage, entrapment, or tip overmight occur, but lack of drive powerwillnotbeafailuremode.

Steering can be accomplished through additional mechanism, but lightweightminimalistapproachesrequirenoadditionalmechanism.Theconfigurationdevisedrotates by skid steering, or turning wheels on one side forward and the otherbackwards.Thisreducesrequiredmechanismmassandcomplexity.Middaytemperaturesatthelunarequatorwouldoverheatmotorsatthewheels,sodrivelines are preferred that remove actuators from wheel hubs. This leads toshoulderdrivenmobilityconfigurations.Shoulderdriveshaveacollateraladvantageof enabling skid‐steer rovers to operate with a single actuator per side. Eachactuatortransmitstorquetobothwheelsononeside.Thisenablestwoactuatorstopropel and steer the rover. Shoulderdrives can transmit torque through seriesofmitergearsanddriveshaftsfromshouldertowheel.Suchsystemshavehighlevelsof complexity, requiring high‐precision gear meshing and complicatedimplementationtoproduceadequatetorque,asthetorquecapabilityofmitergearsisminimal.To reduce complexity andmass and improve torque performance, a chain drivenconfigurationisoptimal.Thissystemlocatesgearreeducationsattheshoulderandtransmits torque to twowheels through roller chain. Roller chain transmits hightorqueswithminimalmassandcomplexity.Suchamechanismhasnotbeenusedinspaceapplicationstodatebecauseofaconcernforcold‐welding.Cold‐weldingisaphenomenon where similar metals fusetogether in a vacuum environmentespeciallywhen forcefully pressed togetheror under thermal swings. This occursbecause of the lack of an oxide coating onmetals in an oxygen free environment. Forshort duration missions, this can beovercome with light coatings of lubricant.Dissimilarmetals canalsobeused in rollerchain elements to prevent cold welding insystems for long duration use. Initialvacuum chamber tests showed no sign ofcold‐weldinginaweek‐longtest.

Figure28:Mitergearsidearmconfigurationtransmitstorquetowheels

throughacomplexgearingsystem.

32

Figure29:Mobilitysidearmdrivestwowheelswithasingleactuatorattheshoulderandroller

chaintotransmittorquetowheels.

Figure29:Testingrollerchaininvacuum.

Iteration of mechanism sought to drive down mass and size of componentsoptimizing based on loads experienced in testing. The result is a lowmass drivesystem.

Figure30:Mobilitymassreducedthroughanalysisofloadconditions.

33

A stack of a brushlessmotor, planetary gearing, and a harmonic drive1 performsshoulder actuation of roller chain in the XPrize rover. Brushless motors areemployeddue to excellent reliability in space environments. Brushlessmotors donothavesparkingproblemscommonwithbrushedmotorsinvacuum,andrejectallgenerated heat through the external casing rather than the rotating portionsimplifying thermal control. Lunokhod used sealed compartments for brushedmotors to overcome this failure mode, and the atmosphere of Mars enabledSojourner to use brushedmotors. A harmonic drive provides themajority of thegearreductionfortheXPrizerover.Harmonicdrivesprovidelargereductionsinasinglestageandhandlelargetorqueoutputsinasmallvolume,lowmasspackage.

Figure31:Harmonicdrivesprovidelargegearreductionsinasmallvolume,lowmasspackage

5.2.2. DriveSuspensionDrivesuspensionseekstomaintainwheelcontactandequalizeroverweightcarriedby each wheel. Equal weight distribution balances the traction potential of eachwheel so that if onewheel encounters an obstacle, the other three are capable ofsupplyingtractiontoassist insurmountingtheobstacle.Equalweightdistributionreduces thedangerof a singlewheelbecomingentrappedby sinking into the soilmorethantheothers.Two sidearms passively rotate about theshoulders and are connected through arocker‐bogie suspension, similar to thatused by Sojourner. This set of passivelinkageswithnoelasticelementsensuresallfour wheels maintain contact with theground.Thislowmasssolutionensuresnearequal weight distribution over all fourwheels on uneven terrain, dramaticallyimprovingmobilitycapability.1TheXPrizeroverusesaMaxonbrushlessandplanetarygearhead[32]coupledwithanHDharmonicdrive[33].

Figure32:Rocker­bogiesuspensionsystem

34

5.2.3. TractionThefundamental functionofamobilitysystemis togeneratetractiontomovetherover. Traction is the result of thrust at the interface between a wheel and theground[28].Tractionforroversisdeterminedbygroundpressureandsoilinterfacedesign. Soil interfacesmustmaintain tractionwhile traveling over loosely packedlunarregolith.Failuresoftractionoccurwhenwheelssliprelativetothetoplayerofsoil,orwhenshearforceswithinthesoilcauseshearfailureswithinthesoil.Successfulwheel surface design grips the soil preventing slip betweenwheel andsoil. Designswith ideal traction do not slip until soil shear failures occur. Roversutilizegrousers,orbladesonthewheelstodigintothesoil increasingthetractiveforcesbyspreadingforcesdeeperintothesoil.Sharpgrouserpointsalsofacilitategriponrockyterrain.

Figure33:Wheelgrousersdigintothesoilincreasingtractiveforces[29]

Decreasing ground pressure reduces soil shear failures by spreading out shearforcesoverlargerareasofsoil.Lowmassandlowlunargravityresultinlowvehicleweight.Numberandsizeofwheelsdeterminecontactpressurewiththeground.Alowweightsystemrequireslesswheelcontactareathanaheavymachinetoreachthe same ground contact pressure. Design for lightweight rovers takes this intoaccount selecting a four‐wheel system tominimizemechanismmass. Variance inwheel diameter and width tune ground pressure by changing contact area. Thisleadstoasystemwithfourlargewheels.

5.3. CameraMechanismsSpace missions require mechanisms to point, place, or manipulate sensors and payload. The XPrize mission necessitates mechanisms to point camera systems and zoom optics.

Camera pointing and zooming require accurate absolute positioning, driving the use of stepper motor actuation. Stepper motors “step” through fractions of a rotation by alternately activating four electromagnets.

35

Figure34:Steppermotorsmaintainabsolutepositioningbysteppingthroughfractionsof

rotationsbyalternatelyactivatingfourelectromagnets.

Cameras are placed high on the XPrize rover and utilize mechanisms to perform panning and tilting motions. Cable management drives designs with cable slack for flexibility during rotations. Panning motions make use of the long distance cable run up a central mast tube to provide slack for cables to twist. Tilting motion bends and unbends cables. Cables are locked in place at either end of the tilt motion to create a region for bending with slack to limit torque requirement of tilt actuation.

Motors to produce pan and tilt motions must be cooled during operation. Motors are located near existing thermal pathways to radiators minimizing required mass for thermal control.

For operations on the moon, telephoto lensing is essential for sensing and data acquisition. Terrestrial telephoto lenses are highly automated and advanced, but no automated telephoto lens has operated on the moon. A lunar telephoto zoom lens for equatorial rovers must be robust to vacuum, radiation and regolith contamination on the lunar surface. To reduce costs development time, rovers make use of existing lens designed for terrestrial application. Modifications are required to replace lubrication and actuators with space-relevant equivalents.

5.4. RegolithProtectionMechanisms must be protected from impingement of highly abrasive lunar regolith to prevent increased friction and eventual failure of moving parts.

During landing, a significant amount of regolith is kicked up by thrusters. Conjoined lander-rovers use the pallet lander structure to shield dust from reaching mechanisms, cameras, and solar/radiator surfaces. Since the moon is a vacuum, all dust kicked up is propelled ballistically, meaning that a small lander pallet is capable of completely preventing dust from reaching the rest of the exploration system.

Figure35:Longdistancefromroverchassistocameraboxfacilitatescablemanagementforpanningmotionsbytwistingcablesinsidethemasttube.

36

Drive mechanisms make use of shoulder drive to shield actuators from regolith. Shoulder actuators have no external rotating parts. Roller chain running from a shoulder to wheels is encapsulated within structural tubes to shield the chain from regolith. Rotations at the wheel are vulnerable to dust intrusion and make use of labyrinth seals to minimize dust reaching internal bearings and chain. Safety factors in drive actuation take up any additional friction caused by dust.

Additional mechanisms for pointing and zooming are located within the chassis or high on the rover where dust is unlikely to reach in significant quantities.

37

6. StructureMinimizingthestructuralmassofspacerobotsmatters,sinceeverykilogramspenton structure cannot be spent on payload like electronics and sensors. Conjoinedlander‐roversdemandtheroboticstructuretobehighperformance,yetlightweight.

6.1. StructuralLoadingThe structure of an exploration systemmust survive a variety of load conditionsthrough the course of a mission. Structures for rover, lander, and connectionsbetweenrocketmotorsmustbeconfiguredrelativetotheseloads.

Accelerations are imparted when rockets fire to launch, inject, or brake. TheseaccelerationsactmuchlikeincreasedgravitationalaccelerationimpartingloadsonstructuresasthoughcomponentswereseveraltimesheavierthantheyareonEarth.Theseaccelerationsarelargeronalowmasssystem,sincesimilarrocketforceswillresultinhigheraccelerationassystemmassdecreases.Structuremustwithstandvibrationsimpartedduringlaunchandstageseparationsandfromacousticeffectsastherockettakesoff.Structuresmustbestifftosurvivevibrationswithminimaldeflections.

In addition, landing impact and driving also impart structural loading. Landingshockabsorptionmustbedesignedtolimitlandingloadsexperienced.Locomotionmustbeintegratedtothechassissuchthatloadsarehandled.

Figure36:Thestructureofaconjoinedlander­rovermustsurvivevibrationandacceleration

loadingduringlaunch,lunarinjection,braking,andlanding.

38

6.2. CompositeMaterialsMuch as the application of advanced magnesium and titanium alloys led to thesuccess of the Lunokhods, the application of cutting‐edge composite materialsprovides the potential to dramatically improve performance of modern lander‐rovers.

Useof fiber‐reinforcedcompositematerials isprevalent in theaerospace industryand composites are materials of choice for today’s cutting‐edge aircraft andspacecraft.Advancedcompositematerials,oncethedomainofgovernmentaircraft,are now ubiquitous. They are utilized in golf clubs, sail and motor poweredwatercraft, automotive race vehicles, production sports cars, wind turbines,commercial jet liners, and structural reinforcement in bridges. Worldwidecompositegrowthisburgeoningintoday’sefficiency‐driveneconomy,usheringinanew era of materials. Utilization of composites has not yet permeated to spacerobotics.Composite structuresunlockanewrealmofpossibility inadvancedspace roboticapplications.Massreductionsofupto30%andincreasesofthreetimesthestiffnesscomparedtotraditionalmetallicstructurescanberealized.Structuralmasssavingsenable unprecedented sensing and actuation payloads. Appendages such as armsandwheelscanbelightened,reducingpowerrequiredforactuation.Stiffcompositestructures survive vibrations during launch and increase sensing accuracy byreducinguncertaintycausedbymaterialdeflections.MaterialpropertyadvantagesofcarbonfibercompositesovertraditionalmetalsareshowninFigure37.

Figure37:Comparisonofcarbonfibermaterialpropertiestometalsrelativetocomponentmass

39

Composite construction layers dissimilar materials to create a “layup” thatharnesses the strengths of each layered material. These layers are placed andorientedoptimallytoaddstrengthandstiffnessintheexactlocationsanddirectionsneeded.Fiber layers, like carbon fiber, used in composite layups are fabricwith all fiberspointedinthesamedirection.Theseareorientedforhighstiffnessandstrengthinthe direction of the fibers. This is unlike metals, which always have the samestrengthandstiffnessineverydirection.Theresultissignificantmassreductionforthesameperformanceinapartbyaligningfiberswithloadpaths.

Lightweight core materials, like aluminum honeycomb, add thickness todramaticallyincreaseflexurestrengthatminimalmass.Complexsurfacegeometrycanbeconstructedsandwichingalayerofaluminumhoneycombbetweenlayersofcarbonfibers.Thiscreatesastructurethatcarriestensionloadsthroughthecarbonfacesheetsandresistsbendingthroughthehoneycombcore.

Figure38:Compositedevelopmentcycle.

6.2.1. RoverChassisTheXPrizeroverusesachassiswithacarbon‐honeycombsandwichskin.Theloadsimparted to the spacecraft are carried through this skinminimizing the need forseparate structural members. This method of carrying loads through the skin isknown as monocoque design. This technique minimizes mass by carrying loadsthrougheverypartofthechassiswithnoadditionalmassforconnectionsbetweenstructuralmembers.

40

Figure39:Carbon­honeycombchassisdesignmakesuseofcarbonfacesheetstohandletensile

loadsandaluminumhoneycombtoresistbending.

The chassis takes full advantage of required surfaces, using carbon‐honeycombpanels to provide backing for solar panels and as the radiator surface. Theseelementsareintegratedthroughacarbonhoneycombbaseandsidewallssothattheentirestructureassistsinhandlingloads.

6.2.2. LanderStructureLander structure must endure loads from upper stages, landing propulsion, andlanding legs.During flightand landing, loadsaretransmittedfromlandertoroverthroughaseparationmechanism.The great configuration problem is to keepthe lander pallet thin. This reduces stackheight in the launch fairing and reduces thestep‐offfortherovertodismount.Successfulconfiguration is accomplished by protrudingthebrakingsolidrocketmotorintothecenterof the lander pallet with propellant andoxidizertanksclusteredcloselyaroundit.The XPrize lander pallet is comprised of atriangular carbon composite shell withstiffenersforstrengthandrigidity.Thispalletis dual‐used as a shield to prevent dustkicked up during landing from reaching therover.

Figure40:Thebrakingsolidrocketprotrudesintothelanderpalletreducing

stackheight.

41

Figure41:Thelanderpalletiscomprisedofatriangularcarboncompositeshellwithstiffeners

forstrengthandrigidity.

6.2.3. SecondaryStructuresSecondarystructureslikelanderlegs,wheels,tubes,andI‐beamsexploitcompositedesigntoreducemass.

IntheXPrizelander‐rover,tubescarryloadsfromthelanderlegs,roverwheelsandcameramechanisms to the chassis. Structural I‐beams inside the chassis providedirecttransferofloadsfromtheradiatorsurfacetothechassisfloor.Theseareusedtosupportheavyelectroniccomponentsthataremounteddirectlytotheradiator.Specialized composite designs have been developed to fit these applicationswithsignificantmassreductions.SamplestructuresareshowninFigure42.

Figure42:Secondarylander­roverstructures,likeI­beamsandtubesexploitcompositedesign

todramaticallyreducemass.

Large wheels dramatically improvemobility performance for a lowmass lander‐rover,butcomeatahighermass.Useofcarboncompositematerialsenableslargerwheelsbydramaticallyreducingmassrelativetometalalternatives.

42

Figure43:Carbonwheeldesignsenabledramaticmassreductionrelativetometalwheels2.

6.2.4. CompositeThermalRegulation

Specialized tyes carbon fiber provide excellent thermal conductivity properties.CompositesareheavelyusedintheXPrizelander‐roverthermalsystemstoprovidehighlydirectionalheatconductionatalowmass.The radiator used to configure a low mass exploration system is a carbonhoneycomb structure. It has highly conductive carbon facesheets, with a densehoneycombcore.Thefacesheetsspreadouttheheatacrossthepanelandthedensehoneycomb increases conductivity through the panel. Honeycomb also addsstiffnesstothepanelensuringgoodcontactbetweencomponentsandtheradiator.Highheatproducingcomponentsaremounteddirectlytotheradiator.Themotorsand Inertial Measurement Unit are thermally strapped to the radiator throughthermally conductive structural I‐beams. These pathways use directional carbonfibertoconductheatefficiently.

Thermal straps and I‐beams are separated from the chassis floor with isolationcones. Isolation cones are made ofextremely low conductive plastic and theconeshapeminimizesconductionwiththechassis, which is 120°C near lunar noon,whilemaintainingstructuralneeds.Solarpanelsoperateathigherefficiencyatcooler temperatures. Thermallyconductive carbon is used as the backingfor solar panels to improve performance.This is accomplished by distributing heataway from illuminated sections andrejecting it to black space from non‐illuminatedsections.

2Thecompositewheelpicturedachievedamassof300gramscomparedtometaldesignswithmassofapproximatelyonekilogram.

Figure44:Highconductivitycarbonfibersareusedtocreatehighlyefficient,lowmass

thermalpathways.

43

7. AvionicsAvionicsaretheelectronicsofaspacesystem.Lander‐roveravionicsmustsucceedwith lowmass and packing volume. Conjoined lander‐rovers dramatically reducemass by eliminating redundancies in avionics. The same avionics are utilized incruise,landing,androving.Allavionichardwareresidesintherover,andtheroveravionics control the landerwhile in flight through electrical umbilical connectorsbetweenlanderandrover.Theseumbilicalconnectorsseparateuponlanding.

Avionics differ from terrestrial electronics in terms of power, thermal control,radiationhardening,andshockandvibrationsurvival.Power for space electronics is limited to minimize requirements for powergenerationaswellastominimizeheatproduction.Heatgeneratedbycomponentsmustbeconductedaway from individual chips in theelectronics to radiators thatreject heat to black space. Power and thermal requirements of avionics drive theoverall size and mass of lander‐rovers by driving the size of solar panels andradiators.Thiscalls for lightweightsystemstoselectavionics thatrequireas littlepoweraspossibletocompleteamission.

Radiation dosage received by the electronics often limits components that can beselected. The Earth's magnetic field protects terrestrial electronics from themajorityoftheradiation.However,onceaspacecrafttravelsbeyondtheVanAllenBelt, it is bombarded by large amounts of radiation. Electronic systems do nottolerate high dosages of radiation because their fundamental building blocks,transistors, are small, energy sensitive devices. Three major types of radiationeventscandisruptadigitalsystem.ThefirstisaSingleEventUpset(SEU)whereahighlychargedioncancauseabittoflip.Theseeventsareeasilyhandledbywritingsoftwarethatcanrecover fromsmallcorruptionerrors.Thenext isaSingleEventLatch‐up (SEL) where a transistor becomes shorted until it is power cycled.Circuitry that is considered to be latch‐up immune includes extra circuits thatautomaticallypowercyclethetransistorswhenanSELisdetected.ThelasttypeofeventisaSingleEventGateRupture(SEGR).Thisoccurswhenahighenergy,heavyparticlepassesthroughtheelectronicsandmakesaholeinthetransistor.Afteranevent like this, the transistor will be permanently destroyed. Gate rupture isextremelyrare.Radiation is a significant issue for spacecrafts that are designed to last formanyyearsandwillreceivelargeradiationdosages.However,forshort‐durationmissionsthe expected radiationdosage isminimal3. Components that arenothardened forradiationcanoftensucceedinthesescenarios.

Duringlaunchandwhenfiringsecondarystages,aspacecraftundergoeshighshocksandvibrations.Undertheseconditions,mostEarthelectronicsfail.

3Aonemonthmissiontothemoonwouldreceivearadiationdosageonorderof1krad[36]

44

7.1. CommercialOff‐The‐Self(COTS)AvionicsTraditional space avionic systems are heavily overdesigned relative to expectedneeds.Mass isaddedtoacomputerchassisto improvedynamic loading,radiationshielding,andthermalmass.Boardsarelooselypackedwithlargechipstoimproveradiationtolerance.Inatraditionalspacesystemlikeasatellite,themassofavionicsisasmallfractionoftotalsystemmasssotheadditionalmassdoesnotsignificantlyimpact launch requirements. Lowmass systems that are sent farther from Earthpush the margins on avionic mass as additional kilograms dramatically impactlaunchrequirements.Conventionally, avionic systems involve a significant amount of customdevelopment and are comprised of special computer chips designed towithstandradiation.Customdevelopmentsimposesignificantcostindevelopmentandtesting.Useof commercialoff‐the‐shelf (COTS) components candramatically reducemassand development cost. COTS components are terrestrial electronicswithminimalcustomization to enable space flight. The avionics of a COTS system still enablescustomizations. Cards with fixed sets of chips are used as the building blocks ofCOTSavionicsrelativetoexchanging individualchips intraditionalavionicdesign.This leveragesconfigurationsthathavebeenimplemented inthepast tominimizedevelopment and testing. COTS systems selected for the XPrize use avionic cardswith higher feature‐density due to tighter packing and smaller individual boardcomponents,reducingthenumberofcardsrequired.

Use of Field‐Programmable Gate Arrays (FPGAs) enables a COTS system to beflexible to a variety of applications. FPGA chips are hardware that can bereprogrammedtoservedifferentfunctionsthatwouldtypicallyrequireexchangingcustomcomputerchips.Forexample,theI/OboardcanbereprogrammedtoacceptRS422,serial,orAnalog/Digitalinputswithoutchanginghardware.FPGAsarealsoused to run specialized processes without overloading the main processor. Forexample,aspecializedFPGAcancompressincomingvideoandimagesautomaticallywithoutloadingtheprocessor.

Mass reductions for aCOTSavionic systemcome largely fromoptimizationof theavionicschassis.A lowmasssystemdoesnothave the luxuryof largemarginsonthermal mass and radiation shielding. Characterization of the expected radiationenvironment based on mission length and destination and the expectation ofthermal swings can be used to optimize themass of the avionics chassis. Chassismasscanbereducedby50%byreducingthesemargins4.

4BasedontheBroadReachEngineeringchassis[34]

45

Figure45:LowmassCOTSavionicchassis

COTS solutions also provide significant benefit in lander‐rover communicationhardware. Ground system compatibility restricts communication signals to lowfrequenciesintheS‐band.Whilethisdoesresultindecreasedbandwidth,S‐bandissimilar to cellular telephone frequencies that have been developed and testedextensively. Cutting edge COTS communication systems replace heavy inefficientdesigns5.

7.2. PoseSensingLocalization is a fundamental problem for anymobile robotic system. In route tothemoonandonthelunarsurface,conjoined‐landersmustdeterminepositionandorientationintheabsenceofterrestrialspecificsolutionslikeGPSandcompasses.Landing requires fast feedback of precise position and orientation to accuratelycontrol thrustersandguidea lander toasafe landingat the intendedsite. InertialMeasurement Units (IMUs) are used to determine acceleration and to derivechanges in position and orientation orpose. Landing radar returns distancemeasurements to determine distance tothesurface.Pinpoint landingrequiresdeterminationof absolute position. Digital SceneMatchingAreaCorrelation(DSMAC) isaterrestrial solution to this problem,commonly found in cruise missiles.DSMACusesacameratocollectimageryduring flight and correlates those withstored satellite imagery to determineabsolute location [9]. Conjoined lander‐

5AeroAstroGalliumNitridebasedamplifiersdoublepowerefficiencyovertraditionalcommunicationavionics.[36]

Figure46:Imagescapturedbylander­roversarecorrelatedwithsatelliteimageryto

determineabsolutepositioning.

46

rovers dual‐use rover navigation cameras to take images required for DSMACeliminatingtheneedforspecializedhardware.Duringsurfaceoperations,IMUandwheelencodersareusedtodeterminemotions.Encoderscountrevolutionstoestimateodometry.ThesameIMUforlandingisusedfor roving. Since fast feedback and high accuracy are required for landing, dual‐using the landing IMU enhances rover capability relative to typical roving IMUs.Whileroving,theimageryrecordedfromnavigationcamerasisexploitedtoperformfeature and structure based registration with satellite maps. Features such asintensity, texture, and elevation can be used to register what a robot sees withorbital maps to localize the exploration system. This is similar to the DSMACapproachforlanding.

Once localizedon thesurface,3Ddata fromstereocameras is incorporated intoadigital elevation map. Techniques such as evidence grids are employed toprobabilistically fuse data from the stereo cameras, the zoom cameras, and theoriginalorbitaldata.Datafusiontechniquesenablethecontinualintegrationofdatafrommultiplesourcestoimproveamap.Thismapisusedtodeterminesaferoutesoftravel.

7.3. MotorControlMotion control is pervasive and essential in all robotics. Traditional commercialmotioncontroliscommon,butitdoesnotexploittechnologytominiaturize,harden,achieve reliability, and do more with less. Lightweight conjoined lander‐roversrequirecustomizedmotorcontroltoreducemassandimprovepowerefficiency.

Field‐ProgrammableGateArrays (FPGAs) enable sensing, commutation, anddrivemotion with the gamut of controlstrategies from classic PID to adaptiveforcemethods.FPGAchipsenablequickdevelopment and customization ofmotioncontrolboards.

FPGA‐based motion control providesflexibility innumberandtypeofmotorsfromsteppermotorstophasedbrushlessmotors. The system is capable ofmonitoring all motors for heart rate,efficiency,andheat.Vastcontroloptionsare available for each of the motors(speed, steps, shutdown, startup,monitor,etc).

7.4. SoftwareSoftware for a conjoined lander‐rover requires specialized development. Sharingtheresourceofasingleavionicsystemdictatessoftwarethatprovidesfunctionalityforcruise,landing,roving,anddatamanagement.

Figure47:CustommotorcontrolboardutilizesFPGAstominimizecomponentry.

47

Landingandrovingsoftwaresharesomebasicfunctionalityandaredistinctinotherfunctions. Solutions requiring software that controls either lander or roverfunctionality leave state uncertainty and add complexity for implementation andtesting.Theapproachforhandlingthisuncertaintythat is introducedinconjoinedsystems is to create two distinct bootable software images, one for cruise andlandingandoneforroving.Switchingbetweenthetworequiresarebootthatresetsthe software’s initial state. Rebooting is a common occurrence in a space avionicsystemasitistheprimarywaytohandledisruptionsduetoradiationevents.Thisapproach speeds parallel software development and testing of lander and roverfunctions.Many low level blocks of code would be duplicated in these distinct images. Tohandle this,acommoncodebank is introducedcontainingcodethatbothsystemscanaccess,likehardwaredrivers.Thisreducesduplicationofcodedevelopmentandenablesbothlanderandrovertestingtocatchbugsinsharedsoftware.

Selectingtheproperoperatingsystemisimportantinaspacesystem.Functionsoflandingrequireprecisetimingdictatingtheneedforareal‐timeoperatingsystem.Possibilitiesdoexistforterrestrialoperatingsystemstobeappliedforspaceflightatlittletonocostforlicensing.Howeverin‐housedevelopmentincreaseswiththeseoperating systems leading to similar cost relative to licensing of space‐testedsoftware.Heritagesoftwareisselectedforalowcostmissionforhighreliabilityandexistenceofhardwaredriversforspaceelectronicstoreducedevelopmenttime6.

6VXWorkswasidentifiedastheoperatingsystemofchoiceduetoflightheritage[38]

48

8. ConclusionRobotic exploration on themoon’s equator from extremely small launch vehiclesdrives the configuration of conjoined lander‐roverswith asymmetricmorphology,chaindrive, skidsteer, innovativestructuraldesignwithcompositematerials, andcommercial‐off‐the‐shelfavionics.

8.1. Claims

8.1.1. Configuration

Theconfigurationofa lowmass,capable lander‐rover forexplorationof the lunarequatorhasbeendevised.Thesystemdevisedconjoinsthefunctionsoflandersandroversreducingredundanciesofstructure,power,computingandcommunication.These payoffs equate to less mass, simplified integration and lower costs thanachievablewithconventionalconfigurations.

Reaching 120kg systemswith conjoined lander‐rovers opens the door to reducedmissioncostthroughtheuseofsmalllaunches.Newambitionsforlunarexplorationcall for fiscal responsibility and safety. Lowmass configurations enable low costexplorationforscoutingandpreparingsitesforfuturehumancolonization.

8.1.2. Morphology

Themorphology of lander‐rovers to explore the lunar equator has been devised.The form of a system is laid out that delivers capable exploration during thechangingsunangles,highsurface temperatures, roughnessof terrain,andupwardangletoEarthatthelunarequatorwithintheconstraintsofsmallsizeandlowmass.

Power and thermal design optimizes equatorial operation. An asymmetricmorphology rejectsheat toblack skyonone side andgathers solarpoweron theother. Tacking like a sailboat enables this low mass unactuated configuration tosucceedinthevaryingsunanglesandextremetemperature.

Cameraconfigurationisminimizedtoreducemass.Thecameraarrangementplacesall cameras high on the rovermast, enhancing operator perception and enablingimagerytobecapturedonanysideoftheroverwithoutredundancy.

Rovercontrol is improvedbypresentingrover interfacesthataresimple,effectiveandsafesuchthat,withminimaltraining,operatorscandirectlycontroltherover.Themajoradvancementisthedesignandintegrationoftoolsforpredictingvehiclesafetyinresponsetocommandsissuedbyoperators.The Earth maintains a near constant position in the equatorial sky, enablingsuccessful communication with a fixed antenna and a set transmission beam.Cutting‐edge antenna technologyoptimizes communicationbandwidth for a givenmission. This system continually transmits data for operation interspersed withimageryandvideo.

49

8.1.3. Mechanisms

Mechanisms with low mass and volume have been devised to survive the harsh lunar environment. The extreme temperatures, hard vacuum, and abrasive dust at the lunar equator drove the configuration of a shoulder driven rover with roller chain, precise camera pointing, and a zoom lens.

Thefour‐wheel,skid‐steerdrivedevisedminimizesmechanismmass,andprovidestractionthroughwheelgrousers.Groundpressureistunedbyvaryingwheelwidthand diameter. Midday operation at the lunar equator presents extreme temperatures mitigated by locating actuators near radiator surfaces. Drive motors are shifted inside the chassis at the shoulders with roller chain to wheels. Transmission of torque drove the innovation of a space roller chain mechanism. This low mass, high efficiency mechanism makes use of specialized lubricants to function in the hard vacuum and extreme temperatures.

Camera pointing and zoom mechanisms developed enable lander-rovers to precisely point and zoom in to provide desired imagery. Thermal and wiring design are specialized for camera pointing mechanisms. Tolerance and lubrication design are specialized for zoom mechanisms.

8.1.4. StructureInnovativestructuraldesignforaconjoinedlander‐roverhasbeendevised.Ahighperformanceroboticstructurewasconceivedrelativetolowmass.

The application of cutting‐edge composite materials provides improvedperformance formodern lander‐rovers.Mass reductionsand increases in stiffnessover traditional metallic structures have been achieved. Structural mass savingsenable unprecedented sensing and actuation payloads. Stiff composite structuresenable survival under vibrations during launch. Structural loads are transmittedthroughthechassisskineliminatingmassofadditionalstructuralmembers.

8.1.5. AvionicsAvionic systems for lightweight conjoined lander-rovers have been devised. Avionics devised optimize mass by reducing margins on power, thermal control, radiationhardening,andshockandvibrationsurvivalrelativetotraditionalavionics.

COTScomponentsareleveragedtodramaticallyreducemassanddevelopmentcost.Avioniccardsusedhavehighfeature‐densityreducingthenumberrequired.UseofFPGAsenablesflexibilitywithoutexchangingchips.Avionicchassisdesignoptimizesmass by reducingmargin on thermal and radiation shielding. Characterization ofexpectedradiationenvironmentbasedonmissionlengthandexpectationofthermalswingsareusedtooptimizethemassoftheavionicschassis.

Pose sensing for landing and roving make use of shared sensors minimizingcomponentry in a conjoined system. Similarly, lander and rover softwarearchitecturessharebuildingblocksofcodetostreamlinedevelopmentandtesting.

50

8.2. FutureWorkThis research lays out the considerations of a conjoined lander‐rover and thesubtleties and pragmatics of implementation. Current state‐of‐the‐art results in alander‐rovermass just inside the 120kg requirementwithminimalmassmargin.Furtherreductionsofmasstoincreasethemassmarginona120kgscenarioortodrop to a smaller launch are possible with future technologies, components, andmaterials.

The possibility exists, for ultra light lander‐rovers, to dispensewith radar and todual‐utilize cameras for landing, roving and science. Cameras can be used toprecisely estimate altitude, velocity, pose, and avoid obstacles. This scenario isincreasingly viable with the advent of ultra‐high quality imagery available fromorbitalmissions andprocessing power for pose estimation. The ponderousmassandpowerconsumptionofradarcouldalsobeavoided.

Innovativevisiontechniques likestructure frommotionorstructure fromshadowpresent theopportunity toprovide3D imagerywithonlyonecamera.ThiswoulddispenseofthepowerandmassofadditionalhardwareandreducethesizeofimagedatasenttoEarth.Increasing popularity of higher frequency communication bands present theopportunity for future lander‐rovers to communicate more efficiently. Higherfrequency communication enables significantly higher bandwidth for the samepowerincreasingthesizeofdataalander‐rovercantransmit.High‐speedtraversesbecome increasingly feasible as lander‐rovers are capable of sending back moredata.

Survivingthelunarnightwouldopenthedoortodramaticchangesincapabilityofconjoined lander‐rovers. Equatorial lunar explorerswould be capable ofmultidayjourneys.ThiswouldenableamissionliketheoneenvisionedtoexpandfromavisittoApollo11intoatouroflunarhistoricsites,trekkingacrossthemoontoseeotherApollo hardware, the Lunokhod rovers, NASA Surveyor landers, and Soviet Lunalanders.Surviving theextremecoldofa ‐170°C lunarnight requires innovationofavionics,batteries,andmechanismswithimprovedthermaltolerances.Innovation of enhanced mobility mechanisms presents the opportunity forexploration in increasingly rugged terrain. The development of mechanisms liketrackedlocomotionwouldenablemissionstoincludeshortventuresintocratersinsearch of ice or into lunar lava tubes, which are sites for potential futurecolonization of themoon. Increased traction also enables the development of lowmassworkmachines,capableofworkingthelunarsoiltoprepareasiteforfuturehumancolonization.

51

Figure48:Innovationsinmobilitywouldenablelowmasslunarworkmachines.

52

9. PerspectiveandContributionThe research discussed incorporated the work of a team of faculty, staff, andstudentsatCarnegieMellonUniversity.MyworkonmissionconfigurationdesignsoftheXPrizelander‐roveranddetaileddesign,analysis,prototyping,andtestingoflowmassspacemechanismsandstructurescontributedtothisproject.I mastered developing configurations for space robotic systems. This includedselecting rockets to launch, transit, and land rovers on the moon as well aspartitioningmass among subsystems. Evolution of systemmorphology proceededas a team effort providing me with perspectives on design of systems includinglanding propulsion, solar and thermal, mobility, communication, and cameraplacement.

My contributions included detailingmechanical designs. I led the design of rovermobility mechanism. I iterated designs, analyses, and prototypes. These includedsizing of actuation, analyses to minimize mass, and tests of wheel traction andobstacleperformance.Ialsoworkedonthedetailingofstructuraldesignforlander‐roverspacecraft.Structure‐relatedcontributions includedleadershipofanalysisofcarboncompositedesignssubjecttoaccelerationandvibrationloadsduringlaunch,cruise, and landing as well as loads imparted during landing and roving. Mycontributions included in‐house manufacturing of composite materials creating afulldesigncycletoiteratecomposites.Participating at every level of theXPrizeproject fromhigh‐level systemdesign todetailedsubsystemdesignprovidedmewithabroadperspectiveon thedesignofspaceroboticsystems.

53

10. References

1. Kermurjian,Alexander.(1990).“FromtheMoonRovertotheMarsRover.”Projects:SpaceInformation.ThePlanetarySociety.http://www.planetary.org/programs/projects/space_information/tpr_1990_4_kermurjian.html

2. Gromov,V.Kemurdjian,A.ETAL.(2003)“Lunokhod2–ARetrospectiveGlanceAfter30Years,”Science&TechnologyRoverCompanyLtd(RCL),J.‐St.Co.RussianMobileVehicleEngineeringInstitute(VNIITRANSMASH).http://www.cosis.net/abstracts/EAE03/14528/EAE03‐J‐14528.pdf

3. Christy,Robert.http://www.zarya.info/Diaries/Luna/Luna17.php 4. Wade,Mark.EncyclopediaAstronautica.1997‐2008.

http://www.astronautix.com/5. “TanksontheMoon”TheScienceChannel6. http://www.ri.cmu.edu/pub_files/pub1/berkelman_peter_1995_1/berkelma

n_peter_1995_1.pdf7. Chaikin,Andrew.“TheOtherMoonLandings:TheSovietslostthemoonrace

butwonadramofglorywiththefirstroboticcrafttoroamanotherworld.”Air&SpaceMagazine,March01,2004.http://www.airspacemag.com/space‐exploration/other‐moon.html

8. Moiseev,V.“TitaniuminRussia.”MetalScienceandHeatTreatment.Vol.47.2005.

9. Head,J,Hoppa,G,Gardner,T.etal.“AutonomousLowCostPerscisionLanderforLunarExploration.”LunarandPlanetaryScience.Volume34.2005.http://www.lpi.usra.edu/meetings/lpsc2005/pdf/1471.pdf

10. Williams,David.“MarsPathfinderProjectInformation.”GoddardSpaceFlightCenter.NASA.2005.http://nssdc.gsfc.nasa.gov/planetary/mesur.html

11. “RoverSojourner:Mechanical,Thermal,andMobilitySubsystem.”JetPropulsionLaboratory.NASA.1996.http://mars.jpl.nasa.gov/MPF/rover/mtm.html

12. Stride,Scot.“MarsMicroroverTelecommunication.”JetPropulsionLaboratory.NASA.1998.http://mpfwww.jpl.nasa.gov/MPF/rovercom/

13. Braeunig,RobertA.“RocketPropulsion.”RocketandSpaceTechnology.2007.http://www.braeunig.us/space/propuls.htm

14. Kellog,Larry.“LunarProspector.”AmesresearchCenter.NASA.2001.http://lunar.arc.nasa.gov/

15. Bate,R.R.,Mueller,D.D.,White,J.E.,(1971).FundamentalsofAstrodynamics.DoverPublications,NewYork.

16. “AutomatedAntennaDesign.”EvolvableSystemsGroup.AmesResearchCenter.NASA.http://ti.arc.nasa.gov/projects/esg/research/antenna.htm

17. Gilmore,DavidE.SpacecraftThermalControlHandbook.Volume1:FundamentalTechnologies.SecondEdition.TheAerospacePress.ElSegundo,CA.2002.

54

18. Karam,RobertD.SatelliteThermalControlforSystemsEngineers.ProgressinAstronauticsandAeronautics.Volume181.AmericanInstituteofAeronauticsandAstronauticsInc.Reston,VA.1998.

19. Sarafin,ThomasP.SpacecraftStructuresandMechanisms,FromConcepttoLaunch.SpaceTechnologyLibrary.MicrocosmInc.ElSegundo,CA.FourthEdition.2003.

20. Wijker,J.SpacecraftStructures.Springer.Berlin,Germany.2008.21. Wijker,J.MechanicalVibrationsinSpacecraftDesign.Springer.Berlin,

Germany.2004.22. Young,Anthony.LunarandPlanetaryRovers:theWheelsofApolloandthe

QuestforMars.Springer.PraxisPublishing.Chichester,UK.2007.23. Harvey,Brian.RussianPlanetaryExploration:History,Development,Legacy,

andProspects.Springer.PraxisPublishing.Chichester,UK.2007.24. Observationsofascienceteamduringanadvancedplanetaryrover

prototypefieldtestJ.Glasgow,E.Pudenz,G.Thomas,P.Coppin,N.Cabrol,D.Wettergreen,Mech.&Ind.Eng.,IowaUniv.,IowaCity,IA,USA.

25. Analysisofscienceteamactivitiesduringthe1999MarsokhodRoverFieldExperiment:ImplicationsforautomatedplanetarysurfaceexplorationG.Thomas,M.Reagan,E.A.Bettis,N.Cabrol,A.Rathe,JournalofGeophysicalResearch,Vol.106No.E4,pp7775‐7783,April25,2001.

26. ImprovingHuman‐RobotInteractionforRemoteRobotOperation,Yanco,H.A.Baker,M.Casey,R.Chanler,A.Desai,M.Hestand,D.Keyes,B.Thoren,P.RobotCompetitionandExhibitionAbstract,ProceedingsoftheNationalConferenceonArtificialIntelligence(AAAI‐05),July2005.

27. AnAnalysisoftheIntelligentRoboticsGroup'sExperiencewiththeMarsExplorationRoverMissionK.StubbsandI.Nourbakhshtech.reportCMU­RI­TR­04­45,RoboticsInstitute,CarnegieMellonUniversity,September2004.

28. Apostolopoulos,DimitriosS.“AnalyticalConfigurationofWheeledRoboticLocomotion.”TheRoboticsInstitute.CarnegieMellonUniversity.April2001.

29. Zakrajsek,J.“State‐of‐the‐ArtAssessmentofMechanicalSystemsforFutureLunarSurfaceMobilityPlatforms.”GlennResearchCenter,NASA.2007.

30. “Frangibolt:Non‐ExplosiveActuators”,TiNiAerospaceInc.SanLeandro,CA.2001.http://www.tiniaerospace.com/fbt/frangibolt.html.

31. Diamandis,Peter.“Blastoff!”http://www.diamandis.com/blastoff.html.32. Program07/08.MaxonMotor.www.maxonmotor.com.33. HarmonicDrive.www.harmonicdrive.net.34. Thibodeau,Chad.“IntegratedApproachwithCOTSCreatesRad‐Tolerant

(SBC)forSpace.”MaxwellTechnologies.COTSJournal.Dec2003.35. IntegratedSpacecraftAvionics.BroadReachEngineering.

http://www.broadreachengineering.com/iau.html.

55

36. Holmes‐Siedle,Andrew.Adams,Len.“Handbookofradiationeffects.”OxfordUniversityPress,Edition:2.2002.

37. “X‐BandTransponder.”AeroAstro.www.Aeroastro.com.38. Stumpf,Matthias.“EmbeddedDevelopmentfortheExtremeDemandsof

SpaceExploration.”WindRiverandNASA.2003