Copyright © 2002 Hamilton Sundstrand Space Systems International, Inc.

ABSTRACT

Hamilton Sundstrand Space Systems International(HSSSI) participated with the National Aeronautics andSpace Administration (NASA), the Search forExtraterrestrial Intelligence (SETI) Institute and SimonFraser University in the 2001 field season of NASA’sHaughton-Mars Project (HMP) to study informationtechnologies concepts and hardware systems foradvanced Extravehicular Activity (EVA). The researchwas focused on developing an improved understandingof the uses of the interface in an exploration / fieldscience context. Interface integration withcommunication, navigation and scientific data systems,and the special challenges posed by the expeditionaryenvironment were investigated.

This paper presents a discussion of the field testsystems, test activities and results. Recommendationsfor future, higher fidelity research are included.

INTRODUCTION

Access to, recording, managing, analyzing, sharing andinterpreting information is at the very heart of explorationscience. The unique operational environment of an EVA(Extravehicular Activity) system makes this a specialchallenge to NASA’s mission for the human explorationand development of space (HEDS). In an EVA system,the challenging environment of space - vacuum, thermalextremes, radiation, extreme lighting conditions – arecompounded by system design constraints – weight,volume, power consumption, oxygen compatibility,mobility, dexterity, and tactility restrictions. Together,these challenges have heretofore severely restricted theinformation support and management options for EVAastronauts (Figure 1). NASA has managed this very

successfully during the Apollo missions and spaceshuttle operations through a combination of training andextensive support from the Earth and from IV (Intra-vehicular) personnel using voice communications(Reference 1). However, future, more ambitiousmissions will make these approaches less effective andless desirable. Information transfer and managementrequirements will be increased at the same time thatopportunities for ground support and the effectiveness ofprior training are decreased. Our research is intended tohelp develop a basis for the design and development ofsystems and techniques that will better meet theseneeds.

Figure 1: Current Extravehicular Mobility Unit (EMU) informationinterfaces are limited.

The complexity of information transfer and managementhas grown through NASA’s history as more complexmissions have been undertaken and as robotic elementshave been introduced and used more pervasively. Still,the ability of the EVA crew person to participate directly

2002-01-2312

Investigation of EVA Information Interface Technology in a Mars Analog Arctic Field Science Setting

Michael F. Boucher, Edward Hodgson and Sean K. Murray Hamilton Sundstrand Space Systems International

Pascal Lee SETI Institute and NASA Ames Research Center

Stephen Braham Simon Fraser University

in this process has remained limited and relativelyrudimentary. Information is still provided to the crewperson in real time by verbal communication, writtenchecklists and limited capacity text displays. A far morerobust capability including digital transfer and flexiblepresentation of a variety of data, graphic display ofinformation, and a wide range of data management andcommand interactions is envisioned (Figure 2).Numerous research activities ranging from NASA studiesof intelligent agents (Reference 2), studies of wirelessnetworking technology (Reference 3), and studies ofpotential EVA information interfaces at HamiltonSundstrand and elsewhere (References 4, 5), are aimedat providing these capabilities.

History makes it clear that this will entail difficultcompromises and trade-offs. To arrive at the best totalcapability, these must be made based on a clearunderstanding of the most important tasks and of whatdesign parameters are critical to meeting those needs.Since the systems involved and the functions they willperform are far removed from present and past practice,experience with the use of early development testarticles in relevant settings is vital to achieving thisunderstanding.

STUDY DESCRIPTION

OBJECTIVES

The principal objective of the field research programdescribed in this paper was to improve ourunderstanding of requirements and design optimizationparameters for a future EVA information interface. Wesought to achieve this before committing substantialdevelopment funds by working with pre-prototypehardware and current state-of-the-art software in the fieldscience setting of a NASA Mars analog researchprogram.

There has been a wealth of prior research dealing withthe design of EVA information interfaces (References 6 -8) and with information system needs for field scienceand exploration (References 1, 2, 9). However, thesehave depended on experience derived from systems andcontexts far removed from those realistically expected ina future Mars mission. For example, the Apollo missionswere accomplished with simple EVA informationinterfaces and relied on extensive verbal communicationand support from Earth for information access andmanagement tasks. More recent field studies haveoperated in significantly different contexts and made useof systems that do not reflect the unique interfaceconstraints and requirements for EVA.

Figure 2: A rich set of EVA information interfaces is envisioned to support complex exploration tasks and scenarios.

EVA crew person

Other EVAPersonnel

Robotic systems

Navigation aids

Habitat / vehiclebased humans

Earth basedhumans

Instruments &tools

EVA System InteractionsAdded Interaction Capabilities Under Study

Voice

Data/TextImages

Command DataPhysical ContactDirect Observation

Our research was intended to combine some of theunique constraints of an EVA operating environment withscience objectives and infrastructure representative of aplanetary exploration mission.

Additional objectives of the collaboration includedproviding a user interface that would support theevaluation of communication and field sciencetechniques from the perspective of an EVA crew person.We expected that by working together in this way, someof the most promising avenues of attack and some of themost critical design issues for all of the systems andoperations involved would be identified early. In thisway, greater progress could be made toward thedevelopment of the required integrated missioncapabilities than would be possible through separate,parallel development of systems for later integration at ahigher level of maturity.

HMP NETWORKS AND SYSTEMS

The EVA information systems were operated ascomponents in the evolving HMP experimental planetaryexploration computing and communication infrastructure.Simon Fraser University, the NASA Ames ResearchCenter, and the SETI Institute have been collaboratingunder a series of Canadian systems developmentprograms, working in partnership with correspondingresearch at NASA. The result has been an increasinglydeep understanding of the requirements of fieldexploration communication systems, and the computingrequirements of sophisticated scientific investigations inremote and hostile locations. The 2001 field seasoninformation system research was the first stage in thelatest of these ongoing research projects, a new programcalled MarsCanada. MarsCanada focuses on directsupport of field exploration activities in Canadian Marsanalog locations, such as the Haughton Crater region.

The lead group in the high-speed regional networkingresearch has been the PolyLAB Advanced CollaborativeNetworking Laboratory at Simon Fraser University(SFU), working in collaboration with the CanadianFederal government’s Communications Research Centre(CRC) and Canadian communications company Wi-LANLtd. Research has been funded by the Canadian SpaceAgency (CSA) Space Exploration Program. It is intendedto determine the requirements for information systemstechnologies that support both robotic and humanplanetary surface exploration.

Starting in 1999, a range of technologies have beenintegrated, delivered, and tested in the HMP fieldenvironment. Studies have taken place on range andspeed requirements for data communication, in thecontext of underlying power and space requirements forvarious field systems.

Human planetary exploration demands mobile computingsystems that interface with communications deliveringand receiving voice, data, and potentially video, to andfrom vehicle computing systems, and from systems inplanetary orbit and on Earth. The vehicles can vary incomplexity and size and may be either crewed or robotic.These different applications require different levels ofcommunications availability and quality of service (QoS)parameters. In addition to average bandwidth,parameters such as peak bandwidth, latency, latencyvariation, bandwidth variation, and bit-error andtelemetry, frame-error rate may be significant. Variouscontrol and experiment requirements on the mobileelements determine the required QoS parameters for thecommunications systems.

The environment that will be experienced by moderninformation systems in human lunar and Martian surfaceexploration poses significant issues for communicationscomponents. Large amounts of high quality data must bemoved over great distances. Wireless datacommunications across the exploration region areessential.

Line of site loss and multipathing can produce severeproblems in implementing these regional data networks.Multipathing results when the same signal reaches areceiver by multiple paths as a result of terrain reflectionsand interferes with itself. It is the primary limiting factorin high-speed communication systems, being the onlyproblem that cannot be reduced by changes incommunication frequency or power usage. The moretopographically complex a region, the higher theprobability that both multipath and line of site fading willinterfere with communication to an EVA site. However,the most interesting geological and biological researchcan be expected to take place in the mosttopographically rugged locations, such as in valleys andin close proximity to cliffs and hills. The HMP wirelesscommunications and data networks have been createdto support experiments to better understand theseconditions and improve our ability to deal with them

The goal has been to evolve systems that provide:

• Highly efficient bandwidth utilization to support highdata rate needs with low power consumption.

• Resistance to multipathing in complex terrain.• The ability to support operations where line of site

communications with primary vehicles or spacecommunication segments are difficult or unavailable.

This requires advanced radio modulation techniques anda complex network topology.

In response, the HMP communication system hasevolved into a layered approach consisting of staticspace communication segments linked to regional high-speed advanced radio networking systems, including

static or vehicle-carried regional repeater nodes. Theregional network infrastructure then terminates in simplerlow-power, low weight, local communication linksproviding outbound and inbound network data streams tohighly power-constrained and mass-constrainedelements, such as EVA crews and robotic systems.

In addition to these high-speed systems, lower speeddata is used to provide GPS location services, inemulation of future advanced positioning data systems.Finally, simple VHF and UHF communications are usedto provide long-range safety audio communications.These systems will be integrated into the datacommunications network in the future to provide a single,consistent, and easily maintained flexiblecommunications and networking environment.

HMP computing systems are integrated over thecommunications network simulating anticipatedexploration systems. A field-hardened server supports awide range of wireless network clients including robotics,vehicles and EVA crews. The network protocols usedare designed to tolerate anticipated communicationsdelays for a planetary exploration scenario, and thesystem includes provisions for simulating these delays tosupport experimentation. Experiments with this systemhave included interaction with remote mission controlfacilities and collaborating investigators through satellitelinks as well as regional network operations.

Communication experiments in the 2001 Field Seasonincluded the use of conventional IEEE 802.11b DirectSequence (DS) wireless local area network (LAN)technology to provide medium-scale communicationsfrom the EVA test system to a fixed simulated rover,relayed to a base camp network, and then on to asatellite communication facility via a high-performanceMulticoded Direct Sequence Spread Spectrum (MC-DSSS) radio network link The MC-DSSS system thusprovided the regional networking infrastructure withlower-power local communications being provided by theIEEE 802.11b system.

Although IEEE 802.11b DS was known to provide onlylimited resistance to multipathing, equipment for its usewith available wearable computers was readily availablefor use this summer. As described later in the paper,multipath sensitivity was evident in the field and limitedthe scope and location of test activities. They can besignificantly reduced with more advanced wirelessnetworking systems that will be deployed in the field infuture seasons. With the more advanced systems inplace, we expect that EVA traverse activities will bepossible in geologically and biologically significantlocations, and at great range from base camp facilities.

HMP Context and Operations

The NASA Haughton-Mars Project (HMP) provided thescientific and exploration context for our study. The HMP

is an international interdisciplinary field research projectcentered on the scientific study of the Haughton impactcrater and surrounding terrains, Devon Island, Nunavut,Arctic Canada, viewed as a possible Mars analog.Integration of the information systems research withHMP scientific studies ensured relevance of the resultsto real field science needs.

Mars analogs may be defined as settings in whichenvironmental conditions, geologic features, biologicalattributes, or combinations thereof offer opportunities forcomparisons with possible counterparts on Mars and forsimulations of partial Martian conditions. While climaticand environmental conditions on Devon fall short of theextremes encountered on Mars, aspects such astopography, terrain types, geologic features and theirvariety, and the overall lack of logistical infrastructuresave that brought in by field investigators representunique combinations of analog attributes. Research onthe HMP presents two areas of thrust: Science andExploration Research. Studies are conducted in a multi-disciplinary approach as intimate ties exist between themany disciplines involved.

The HMP Science program seeks to:

1. Compare the Earth and Mars, in particular the manyapparent similarities in geology between DevonIsland and Mars

2. Study the effects of meteorite impacts on the Earth3. Study life on Earth in the extreme environment of the

high Arctic including the preservation and mode ofexposure of biosignatures, and practical issues ofplanetary protection.

The Science program is intended to lead to a betterunderstanding of the nature of Mars and that of the Earthas a planet, of possibilities of finding life on other planets,and of how to go about looking for any evidence of suchlife. These goals are central to NASA’s astrobiology andplanetary exploration programs.

In addition, field science activities on the HMP serve asthe basis on which the HMP’s Exploration Researchprogram is built. This program seeks to use theopportunity of ongoing scientific field research on Devonto develop new technologies, strategies, and experiencein human factors relevant to the future exploration ofMars and of other planets by both robots and humans.

The Exploration Research program is divided into threeareas of research with substantial overlap between them:

1. Information Systems (IS) studies of communicationsarchitectures and information technologies includingstudies in Human-Centered Computing (HCC) andthe development of Mobile Exploration Technologies.

2. Robotic exploration technologies and strategiesincluding field tests of robotic vehicles and

components, definition of the requirements forrobotic systems to support human Mars exploration,and the deployment and evaluation of operationalrobotic systems in response to the exploration andsurvey needs of the HMP Science program.

3. Human Exploration Research investigates manyaspects of field studies in support of the humanexploration of Mars and other destinations in space.Studies to date have addressed exploration logisticsand operations requirements, the planning andexecution of EVAs, EVA systems (e.g., spacesuits,field instruments, tools, roving vehicles, roboticassistants), field science protocols, interactions with"Mission Control" (incorporating 4 to 20-minute timedelays), field medical equipment and telemedicineprocedures, nutrition in exploration environments,habitability, greenhouses, and human physiologicaland psychological monitoring.

The research reported here is interdisciplinary, resultingfrom combining investigations on information systemsand human exploration in the context of supporting fieldgeology studies.

EVA INFORMATION INTERFACE TESTSYSTEMS

SUITED ENVIRONMENT SIMULATOR

Figure 3: Rear view of the simulator showing CPU, optical displaysystem, video converter box, brushless DC motor, batteries, mouse

pad, GPS, and two way radio.

The Suited Environment Simulator (Figure 3) is part of asystem concept mockup designed and built in 1997 toinvestigate an exploration system concept that used afield-replaceable life support system. The hard uppertorso (HUT) was modified for this field season to supporttest integration of a display, computer and controldevices, communications and GPS. A field seasonconsists of several months each summer whenresearchers work on Devon Island at the HMP. Thepolycarbonate bubble assembly simulates the visualdeficits that may occur during planetary explorationoperations and acoustic environments inside a closedspacesuit helmet. The bubble provides impactresistance and in case of emergency can be removedquickly. Two removable arm units were used on thesimulator during this field season. Each arm uses aStereolithography Apparatus (SLA) scye bearing andan SLA upper arm bearing to provide representative armgeometry and deformation of the arm cloth elements,and includes a standard Extravehicular Mobility Unit(EMU) glove interface. The HUT uses a harness systemobtained from a commercially available backpack. Theharness system distributes the system weight toward thehips and can be adjusted to fit a variety of suit subjectsizes. When properly fitted to the suit subject, theharness effectively controls center of gravity (C.G.) shiftsthat might occur during field activity,

The simulator employs a brushless DC fan to provideventilation to the suit subject without generatingelectromagnetic interference (EMI) that would interferewith the test electronics. Although the system is openand ventilation can occur through the open waist bearing,the fan provides up to 36 cubic feet per minute (cfm) ofairflow inside the HUT. A minimum flow of 6 acfm isrequired for safe operation of the suit. A flow diverter ismounted inside the HUT to direct the fan flow over thetop of the test subject’s head. This provides foradequate CO2 washout and moisture removal from theinner bubble surface. For power, the simulatorventilation system shares a 12V, 18 amp-hour lead-acidbattery mounted on the lower portion of the harnessbelow the HUT. An indicator light mounted on the insideof the HUT is continuously illuminated during safe batteryoperations (above 11.5 VDC). Below this voltage, thelight will turn off indicating that it is time to terminate testoperations with the bubble assembly in place. This isillustrated in Figure 4.

Fan

HUT

Battery

Indicator Light

Figure 4. The simulator is configured to represent suited informationinterface operating environments and ensure test subject safety.

3000 Series EMU glove thermal micrometeoroidgarments (TMG’s) were used in conjunction withstandard issue cotton EMU comfort gloves to simulatedexterity and tactile deficiencies associated withpressure suit gloves.

Figure 5: The test system integrates commercially availablecomponents to provide a complete wearable information interface test

bed

Commercially available electronic components wereintegrated with the simulator to support the research asillustrated in Figure 5. All electronic hardware ismounted on the rear of the HUT. An outer coveringencases the electronics and protects the electronicequipment from inadvertent contact. Through the use ofair inlet filters mounted inside the enclosure, some

protection from particle contamination is maintained.The electronics are accessible only when the cover isremoved. However, a simple system was installed tosupport cycling the Central Processing Unit (CPU) onand off, while suited, if a hard reboot was necessary.

Two simulator computer systems were used during thefield tests. One was assembled from discretecomponents using a wearable size CPU with a Pentium3, 800 MHz processor and 256M of RAM. It provideddesirable operational characteristics for speechrecognition system control and display integrationexperiments, but was not fully ruggedized for extendedfield use and could not be adapted for connectivity withthe HMP wireless networks during this field system. Theother unit, an integrated, ruggedized, Xybernaut MA-IVwearable computer, was equipped for networkconnectivity and was used during many field activities.Specific characteristics of each of theses options arediscussed in subsequent paragraphs. Both PC’sprovided serial, PCMCIA and USB ports to interface withvarious types of peripherals. Computer output waspresented to the test subject through a head-mountedmicro-display that could be variably positioned.

Two micro-display formats were studied, a glassesmounted display from Micro-Optical Corporation(Figure 6-1), and the head-mounted Xybernaut opticaldisplay, (Figure 6-2). Comparison and evaluation ofoperational characteristics of these options was aprimary focus of the research.

Figure 6: 1) MicroOptical GMD worn by Dr. Steve Braham (Photocourtesy of Pascal Lee 8061953) and 2) Xybernaut optical display

worn by Trish Garner

During test operations, the subject manipulated thecomputer system using either voice commands or avariety of manual input devices mounted to portions ofthe suit simulator. The speech recognition system usedan ‘off-the-shelf’ microphone headset and plugged intothe computer through the sound card.

One of the manual control interface studies was acommercially available mouse pad system. Laboratoryexperiments confirmed that this device would notrespond to a gloved hand. To solve this difficulty, thepad was enclosed in a typical anti-static metalized mylarbag. (Figure 7) The bag was grounded to provide theelectrical conductivity necessary for gloved operations.This configuration was shown to respond well to a

1 21 2

STAND ALONE COM SYSTEM

(USED WITH XYBERNAUT ONLY)

(18V USED WITH POCKET PC ONLY)

VGA TO HMD CONV.

GLASS MOUNTED

DISPLAY

Battery&CircuitBreaker

12V 18V

FUSE

VENTFAN

BATTERYINDICATOR

HELMET MOUNTED

LED

PORTABLEPC

640X480 VGA

COLOR

SPEAKEROUTPUT

MICINPUT

LOCALBATTERY

IEEE802.11BWIRELESS I/F

DIGITAL

RADIO

RF

SPEAKEROUTPUT

MICINPUT

GPS

LOCALBATTERY

STAND ALONE COM SYSTEM

(USED WITH XYBERNAUT ONLY)

(18V USED WITH POCKET PC ONLY)

VGA TO HMD CONV.

GLASS MOUNTED

DISPLAY

Battery&CircuitBreaker

12V 18V

FUSE

VENTFAN

BATTERYINDICATOR

HELMET MOUNTED

LED

PORTABLEPC

640X480 VGA

COLOR

SPEAKEROUTPUT

MICINPUT

LOCALBATTERY

IEEE802.11BWIRELESS I/F

DIGITAL

RADIO

RF

SPEAKEROUTPUT

MICINPUT

GPS

LOCALBATTERY

(USED WITH XYBERNAUT ONLY)

(18V USED WITH POCKET PC ONLY)

VGA TO HMD CONV.

GLASS MOUNTED

DISPLAY

Battery&CircuitBreaker

12V 18V

FUSE

VENTFAN

BATTERYINDICATOR

HELMET MOUNTED

LED

PORTABLEPC

640X480 VGA

COLOR

SPEAKEROUTPUT

MICINPUT

LOCALBATTERY

IEEE802.11BWIRELESS I/F

DIGITAL

RADIO

RF

SPEAKEROUTPUT

MICINPUT

GPS

LOCALBATTERY

spacesuit gloved hand and provided environmentalprotection for the mouse pad as well. As a result it wasincluded for evaluation in the field study.

Figure 7: Mouse pad enclosed in a metalized Mylar bag.

Voice communications were conducted though aportable two-way radio mounted to the suit simulator.For this study, separate headphone and microphoneinterfaces supported the voice communications andcomputer control voice interactions requiring two devicesto be worn inside the helmet simultaneously for sometests. This compromise simplified hardware preparationfor the tests, but was quite inconvenient during testoperations. Suit subject interface was accomplishedusing the radio’s VOX function or a push to talk (PTT)button soft mounted to the front of the HUT just belowthe bubble interface for accessibility during suitedoperations.

Tracking was conducted through the use of a smallGlobal Positioning System (GPS) receiver. The GPSmade use of a connection to the two-way dual-band(UHF/VHF) radio to transmit a position every fewminutes.

COMPONENT-BASED SYSTEM AND SELECTION

An information display compatible with the physical andoperational constraints of a space suit was perceived asthe most challenging hardware element of the system.Consequently, system development began with thedisplay and proceeded from there to other componentsconsistent with the display and planned test objectives.In addition to operational capabilities, componentselections for the test bed were based primarily on thekey parameters of power, weight, volume andadaptability. Power for the various components had tobe kept to a minimum due to the use of a battery duringthe experiments and the desire to provide for a minimumof four-hour EVA scenarios.

Display Selection

Multiple display options were evaluated before the finalselection of the MicroOptical display for the fieldwork.Attributes of the MicroOptical display which tradedfavorably against the competition were the small size andlight weight of the optics when mounted to the glasses,low power requirements (300mW), ease of use and PCinterfaces.

It provided a full-color, VGA resolution suitable forpresenting the range of text and graphics informationenvisioned in the test program. In the full color mode, itprovides a display brightness of 50fL but can beswitched to a high-bright (240fL) monochrome mode inhigh ambient brightness conditions. The optical displaycan be positioned at various locations with respect to thewearer’s eye providing a 16° field of view and aneffective focal adjustment distance of 2 to 15 ft providingcomfortable display use despite its close proximity to thewearer’s eye.

The display was configured to clamp to the temples ofconventional prescription glasses. Where the user doesnot wear glasses, a pair with non-prescription lensessupplied with the display can be used. These hadtemples that wrapped around the ear to prevent theglasses from moving on the subject’s head during use.Laboratory testing verified that it could beaccommodated within the envelope of a spacesuithelmet without severely restricting the wearer’s headmotions and could be comfortably viewed in positionsthat did not interfere with direct vision for normalactivities. This was considered a key requirement forEVA use.

The display optical unit connects to a VGA conversionbox through a custom pigtail. The VGA conversion boxhas a dedicated power supply and a standard VGAinterface connector, which was connected to the VGAoutput of the test computer. Test system assembly wascomplicated by the fact that the pigtail between thedisplay and the conversion box was of only modestlength and hardwired with no interface connector. Newerversions of the equipment provide a connector, a changethat will significantly aid future research.

PC Selection

The computer for these experiments was intended tosupport delivery of multiple data formats andrepresentative content to the display during field scienceactivities and to permit experiments with a variety ofinterface control concepts. In particular, speechrecognition was envisioned as a particularly attractiveinterface for system control and for field sciencedocumentation. As a result, primary factors in selectingthe computer unit were size, weight and powerconsumption, compatibility with the selected display,

compatibility with candidate control interfaces, adequateprocessing power and memory for speech recognitionsoftware and compatibility with standard software toolsfor display generation and software control.

Various portable PC vendors were evaluated with thefinal selection being a device offered by SaintSong, adivision of Samsung from Korea. The PC provided thebasic attributes required for the field test, VGA formatdisplay, Windows operating system, and an 18 VDCpower interface. Additional benefits were that the floppydisk drive and keyboard were add-ons, which could beremoved during fieldwork. One drawback to the PCselected was the lack of a PCMCIA slot for use with thewireless Internet card. For future work USB wirelessinterfaces will be evaluated. For wireless testing, the PCwas swapped out for the Xybernaut system, which hadthe PCMCIA card capability.

Xybernaut MA-IV Integrated Wearable Computer

Many of the system tests were based on the use of aXybernaut MA-IV wearable computer, providing 300 Mhzof processing power with 192 MB of RAM and 10- GBstorage. These devices support PCMCIA cards,providing a convenient interface for wireless networkingtests. Input/Output can be provided by a range ofdevices, ranging from high-resolution color HeadMounted Displays (HMD), wrist-mounted keyboards anddisplays, and handheld touch-sensitive daylight readableflat panel displays (FPD). Two different head-mounteddisplay configurations were tested with this system asshown in Figure 6. With the use of an I/O expander port,we were able to use the MicroOptical display as analternative to the somewhat larger, although higher-resolution, display provided with the Xybernaut system.

The configuration tested during the HMP 2001 FieldSeason used Microsoft Windows © 98 Second Edition asthe installed operating system base, combined withIBM’s Via Voice Millenium Edition © speech recognitionproduct and the ESRI ArcExplorer © GeographicalInformation System (GIS) product. A range of IEEE802.11b DS PCMCIA cards were used for localcommunications off the suit.

Testing was also performed using a simple vest and/orexternal monitors and keyboards, providing adevelopment and base system test environment thatcould operate independently of the integrated suit mock-up. Testing in the vest included use of advanced server-based voice-driven interface access.

FUNCTIONAL EVALUATION

This year’s field study objectives concentrated mainly onthe human interface requirements for an advancedspacesuit information interface for exploration missions.Testing was performed with a commercially available,wearable CPU and evaluated the ability to present usefulinformation in various formats to an explorer in aspacesuit and the ability to control the informationprocessing and display system in several modes. Theseincluded speech recognition control, manual controlinterfaces and remotely operating the CPU usingconventional wireless LAN technology while in anenvironment that resembled the Martian landscape. Thesimulations evaluated the performance of current helmetmounted displays (HMD), in-helmet mechanical andelectrical integration strategies, software applicationissues, and task interface requirements.

Prior to the start of testing, several persons were chosenbased on technical background, size, age, and gender.The test subjects included both geologists andengineers. Subjects were both male and female andranged in age between approximately 20 and 50. Thephysical differences among the test subjects challengedsystems for establishing and maintaining required headand display positions within the simulator helmet. Thiswas accomplished through the use of an adjustableharness adapted from an internal frame backpack. Aftereach test subject donned the simulator, supportpersonnel adjusted the harness to fit. Harnessadjustment supported proper head positioning andprovided long-term comfort and stability to the subjectduring test operations.

Each subject was asked to develop a voice model usingthe speech recognition software. Voice models weredeveloped while wearing the simulator, with bubble in-place, to account for the acoustic effects of the helmetand air flow environment. Because most test subjectswere not identified or familiarized with the test hardwarein advance, the voice models had to be develped in thefield at Devon Island in a single session for each subject.This proved to be a tedious and uncomfortable process,particularly with the Xybernaut computer in which theslower processor speed resulted in a lengthy post-processing period for the voice model data file. Creatingvoice models over time before going into the field infuture research studies is expected to make better use offield time and to improve the performance of speechrecognition systems.

However, subsequent laboratory tests have shown thatthe need for higher processor speeds was primarilydriven by the use of dictation-driven context-free speechrecognition. Advanced speech recognition tests usingthe Xybernaut vest configuration with a context-driven,speaker-independent voice interface model producedexcellent access to remote data resources, even in high-

noise environments. This is based upon moving theintelligence from brute-force speech model processing toprocess and context-driven processing. The result is lessrequired processing time, speaker and accentindependence, fast recognition time, and sophisticatedhuman-computer interaction. This provides an available,robust command interaction technology, but stilldemands further advances to address the desiredcapability to use the information interface as a flexiblefield documentation aid.

During the field study, the test subjects participated inseveral activities typical of what scientists wouldundertake during a planetary exploration. Activitiesincluded climbing, walking, bending, grasping, roverinterface operations, and tool manipulation whileinterfacing with the simulator’s CPU. The simulator,(HUT, Bubble, Arm Assemblies, Gloves…etc), wasintended to provide a representative environment fordisplay and control interactions (mechanicalinterferences, optical barriers, acoustic environments,etc) while performing these activities. The researchteam recorded feedback from the suit subjects anddocumented events photographically.

The first phase of the research began with testing thesuit CPU, peripherals, including the MicroOptical GMD,mouse pad, and microphones, and optimizing theirintegration with the simulator. This phase of the testingwas conducted close to base camp to make it easy tomake necessary adjustments while debugging thesystem. This phase included evaluation of the voicerecognition software for both field documentation andsystem control. Results for field dictation wereencouraging, but showed the need for greater training ofthe voice model than was practical during the fieldprogram. It was especially notable that the trainingprotocol for the commercial package used did notestablish reliable recognition of scientific terms requiredfor field geology and biology. Speech recognition errorsin these tests were too frequent for reliable display andsystem control, forcing users to rely on a back-up mousepad system.

The mouse pad system worked well with an unglovedhand but the metalized Mylar bag did not providesufficient grounding for consistent mouse operation in thefield making some further refinement necessary. Alayered approach with a sheet of aluminum foilsandwiched between two metalized Mylar bags workedthe best. Several possible mouse pad mountinglocations were also evaluated to identify the bestlocations for this type of manual user interface. Threelocations, the wrist, top of the hand, and front of the HUTwere found to be suitable for effective operation.

Comparative evaluations of mouse pad control use inthese locations (Figure 8) led to the conclusion that awrist or forearm position was best. Users were able to

control the system based on the cursor position on thescreen without direct vision of the mouse pad, but foundthis more difficult. Although the top of the hand seemedto be a good location for the mouse pad it was thoughtthat it could interfere with hand operations in confinedspaces possibly becoming dislodged. The forearm andwrist locations generally offered good control andminimal potential interference. However, this interfacelocation permitted significant changes in relativeorientation of the mouse pad and display axes as a resultof head and arm mobility. At times, cursor movementcontrol was noticeably more difficult as a result.Experiments in this test series were not extensiveenough to determine if this difficulty would besignificantly reduced through learning and experience.This requires further study in the future.

Figure 8: Alternative mounting locations for mouse pad a) top of hand,

b) wrist, c) front of HUT, d) forearm.

Evaluations of the MicroOptical GMD were promising.Because of its small size and mounting location on theeyeglasses, most users felt that normal head movementwithin the confines of the bubble were not obstructed.When correctly positioned, the display providedcomfortable readability without obstructing the wearer’sview of outside objects and task areas. Testing wasperformed in daylight and without the use of a sunvisor.As a result, most users preferred the display in brightmode, which improved viewing while in bright sunlight.Suit subjects commented negatively about the need forreadjustment of the display position during suitedoperations. The positioning adjustment device on thismodel allowed some movement when locked, causing areduction in the view of the screen. This mechanicallinkage must be improved or replaced with anotherdesign for long-term use in a pressurized space suit.

The larger Xybernaut optical display was alsoevaluated. Because of the size restrictions caused bythe helmet bubble, the Xybernaut optical display couldonly be used without the bubble attached. Subjectsfound that the parabolic reflector adequately displayedthe computer screen in all conditions but noticeably

obstructed the wearer’s field of view. Substantialchanges would be required for successful integration intoa spacesuit.

The second phase of test activities focused on wirelessnetwork system integration. The suit simulator wasoutfitted with a Xybernaut MA-IV wearable computer. AIEEE 802.11b DS PCMCIA card with radio antenna wasinserted into the PCMCIA slot and supported the RF LANsignal. A suit-mounted radio antenna provided a radiusof exploration of approximately 400 yards from anintermediate transmitting station intended to represent arover mounted relay station. (Figure 9) The intermediatetransmitting station included another 802.11b wirelessInternet adapter with an additional 802.11b PCMCIA cardwith radio antenna connected to a Linksys wirelessadapter point. High and low frequency parabolicantennas were connected to the Linksys adapter pointand transmitted the LAN signal to the base camp CPU.

Figure 9: Intermediate Transmitting Station

The second phase of testing was done in two stages.The first, preliminary stage focused on establishing andcharacterizing the wireless network links between basecamp computers, mobile relay station, and personalexploration system. The second stage attempted toapply the system in more realistic exploration scenarios.

In the first stage, the transmitting station was assembledon a hill approximately 0.5 km from base camp. The sitewas geologically uninteresting; so, the effortsconcentrated on establishing QoS parameters andenvironmental interferences. If a line of site was keptbetween the transmitting station and the suit computer,the LAN signal was maintained with a range ofapproximately 400 yards. Orientation of the suit subjectrelative to the transmitting station to maintain a direct lineof site between the 802.11b radio cards was required forreliable data communications. This will require furtherwork on antenna integration with the EVA suit in thefuture to avoid constraints on field science activities.

When a repeatable, strong LAN connection wasestablished the second stage of testing commenced. Alocation 1.9 km away from base camp was selectedbased on its distance and geological interest. Severalattempts were made to establish a strong LAN signal, butone could not be established. Signal failure was due tomultipath interference caused by the vertical rock face ofthe prospective exploration site. The relay site wasmoved 0.1 km away from the vertical rock face toeliminate the interference. At the closer site, a strongLAN signal could still not be established. Signal failurehere was also attributed to multipath arising from signalreflection caused by changes in elevation between thebase camp and transmitting station. The transmittingstation was moved to a new location approximately 1.7km from base camp with fewer elevation changes alongthe line of sight. A strong signal was established fromthis site. With a direct line of site between the radiocards, the transmitting station permitted the simulator toestablish a network link as far as 1.9 km from basecamp.

CONCLUSION

This year’s field-testing showed that the simulator couldsupport field test use. Both displays were visible anduseable during exploration activities, however, theXybernaut optical display did not support operationswith the bubble installed on the HUT. Suit subjectsthought that both displays provided adequatepresentation of text and graphics data without specialsoftware, although better resolution and visibility aredesirable. The larger size of the Xybernaut opticaldisplay provided more interference with externalenvironment visibility when compared to the smallerMicroOptical GMD. This field trial shows that anenhanced vision system could potentially be locatedwithin the suit but hardware should be better integratedwith the helmet to minimize helmet strikes during headmaneuvering and eliminate the need for adjustment.

Testing with the wireless LAN system demonstrated thatcommercially available wireless LAN components couldbe assembled in the field and carried the LAN signalnearly 2 km. However, the commercially availableequipment was not designed to operate in the field andwas susceptible to environmental interference. Oneobvious shortcoming of the system was the Loss ofSignal (LOS) caused when the suit subject camebetween the intermediate transmitting station and theIEEE 802.11 PCMCIA radio card. Effective antennaintegration with the suit is needed to eliminate thisproblem. Future field experiments should includeevaluating solutions to this problem that will be robustwithout interfering with mobility. The communicationsystem on the next generation interplanetary explorationsuit could use similar wireless technology, but higherprocessing speed and better software compatibilitycoupled with a larger bandwidth would be necessary to

High and Low Frequency Parabolic Antennas RoamAbout

wireless adapter with RF Card

Linksyswireless adapter point

High and Low Frequency Parabolic Antennas Wireless

adapter with RF Card

Linksyswireless adapter point

High and Low Frequency Parabolic Antennas RoamAbout

wireless adapter with RF Card

Linksyswireless adapter point

High and Low Frequency Parabolic Antennas Wireless

adapter with RF Card

Linksyswireless adapter point

support typical data transfer. Smaller packaging andmore robustness would be important to meet weight andvolume constraints and survive the expected abuse.

FUTURE PLANS

Based on the results of the field-testing described in thispaper, continuing research is planned during 2002 tobuild on and extend what has been learned to date. Thiswill include development of more fully integrated systemswith expanded capabilities, further field tests in thecontext of ongoing scientific research at Devon Island,and expansion of the collaboration to include testing theEVA information interfaces in higher fidelity EVAhardware systems. This may be accomplished bysupporting a NASA JSC field test program using anadvanced EVA prototype pressure suit.

Field tests at Devon Island during 2002 are expected tosupport a more complete and realistic suite of humanscience activities and field science collaboration tests asa result of the implementation of more robust wirelessnetworks and more capable human information interfacesystems. These are expected to enable transfer of thetypes and amounts of data envisioned for ultimate usesof the system and to extend the terrain and distance overwhich testing can be accomplished. More capablehuman interface systems are expected to supporteffective use of system control by the EVA scientist invarious modes envisioned for an ultimate flight designsupporting the evaluation of their usage characteristics ina realistic field research setting. These tests will alsoplay an important role in maturing the human informationinterface hardware, software, and control concepts priorto their use in higher fidelity hardware as part of theNASA JSC-led field test program.

The NASA JSC-led field test program is expected toprovide an opportunity for the evaluation of system fieldperformance in a functioning pressure suit environment.It will also permit the evaluation of robotic systeminteraction applications of the interface in the context ofNASA research on EVA robotic cooperative explorationresearch tests and provide actual experience with longdistance collaboration over satellite data links.

ACKNOWLEDGMENTS

The research described in this paper was supported byNASA, the Canadian Space Agency, and HamiltonSundstrand Corporation. Field research activities wereconducted under the auspices of the NASA Haughton-Mars Project (HMP). The HMP is managed by the SETIInstitute. Logistical support in the field was provided inpart by the Polar Continental Shelf Project (PCSP) ofNatural Resources Canada and by the United StatesMarine Corps (USMC). Permits to carry out research onDevon Island were issued by the Nunavut ResearchInstitute (NRI). The northern communities of Grise Fiordand Resolute Bay are thanked for their overall support ofthe HMP. Xybernaut Solutions, Inc. donated the

Xybernaut wearable computers used in this research tothe SETI Institute. Kawasaki Motors USA sponsored theATVs.

The authors would also like to thank many individualswho supported the research in the field and during thesometimes difficult process of preparing field systemsand getting them to Devon Island. Joel Goldfarb, ClarkDean, Pete Kinsman, and Peg Pozdol of HamiltonSundstrand and Colleen Lenahan of the HMP workedtirelessly to ensure that functioning EVA simulator andinformation interface hardware were available to supportthe research at Devon Island when needed. John Schuttand Gordon Ozinski provided capable and essentialsupport as knowledgeable test subjects in the field. Wewould also like to acknowledge the support of NASApersonnel, notably Richard Fullerton of the EVA ProjectOffice and Joe Kosmo at the Johnson Space Center whoprovided valuable input on desirable interface designcharacteristics, anticipated usage patterns and testingapproaches. Geoffrey Briggs, Brian Glass, and AnthonyGross of the Center for Mars Exploration at NASA AmesResearch Center also provided instrumental support.

REFERENCES

1. Carr, Christopher E., Newman, Dava J., “SupportingMartian Field Geology”, Geological Society ofAmerica, Boston, MA, November, 2001.

2. Clancey, W.J., “Human Exploration Ethnography ofthe Haughton Mars Project – 1998 – 1999”, MarsSociety Annual Meeting, Boulder, CO, 2000.

3. Braham, S.P., Anderson, P., Alena, R., Glass, B.,Lee, P., “Internet Communications for PlanetaryExploration”, Proceedings of the Space internetWorkshop, 2000.

4. Carr, Christopher, E., Newman, Dava J.,“Applications of Wearable Computing to Explorationin Extreme Environments”, MIT, August, 2000.

5. Gernux, Carolyn G., Blaser, Robert W., Marmolejo,Jose, “A Helmet Mounted Display DemonstrationUnit for a Space Station Application”, SAE 891583.

6. Goodman, Gail, “Information requirements for SpaceStation Freedom EVA”, SAE 919526, 1991.

7. Carr, Christopher, E., Schwartz, Steven J., Newman,Dava J., “Preliminary Considerations for WearableComputing in Support of Astronaut ExtravehicularActivity”, MIT, 2001.

8. Simonds, Charles H., Chen, Hsiang Chen, “Designand Testing of An Electronic Extravehicular MobilityUnit (EMU) Cuff Check List”, SAE 911529, 1991.

9. Cohen, Marc M., “Mars Surface Science LaboratoryAccommodations and Operations”, SAE 1999-01-2142, Danvers MA, July, 1999..

CONTACT

For more information on the NASA HMP, visitwww.marsonearth.org

DEFINITIONS, ACRONYMS, ABBREVIATIONS

EVA: Extravehicular Activity

CPU: Central Processing Unit

CSA: Canadian Space Agency

DS: Direct Sequence

EMU: Extravehicular Mobility Unit

ESA: European Space Agency

GMD: Glass Mounted Display

GPS: Global Positioning System

HMD: Helmet Mounted Display

HMP: Haughton-Mars Project

HSSSI: Hamilton Sundstrand Space Systems Intl.

HUT: Hard Upper Torso

IS: Information Systems

JSC: Johnson Space Center

LAN: Local Area Network

LOS: Loss of Signal

NASA: National Aeronautics and Space Administration

PC: Personal Computer

PCSP: Polar Continental Shelf Project

PTT: Push to Talk

QOS: Quality of Signal

SETI Institute: Institute for the Search for ExtraterrestrialIntelligence

SLA: Stereolithography Apparatus