" -l . i . ,.. •,:1 . . .. , , . .. I I .a , . k I_ il

U.S. Army Research Institutefor the Behavioral and Social Sciences

Research Report 1525

N" MANPRINT Support of Aquila, theArmy's Remotely Piloted Vehicle:

O Lessons Learned

John E. Steiprt 11 Edwin _R Smotz, and Nigel R. NicholsonU.S. Army Research Institute

June 1989 -"

Approve- 'or pubi. ;eea..: di-tnbution is unliiii,,ed

U.S. ARMY RESEARCH INSTITUTE

FOR THE BEHAVIORAL AND SOCIAL SCIENCES

A Field Operating Agency Under the Jurisdiction

of the Deputy Chief of Staff for Personnel

EDGAR M. JOHNSON JON W. BLADESTechnical Director COL, IN

Commanding

Technical review byi

Richr.-d E. Maisano -_jJohn L.. .iiies, Jr. -

NOTICES

DIS. TRIBUTION: Pnaydsri olý i on of this repo71 e ade by ARI ditK desscurr ndcncc concý rn itiuti of reports 0:U.S. Ay ani tt.frthi

a

I 'b : a _!ord i ' jrg-

-hla-, al and S(I Sciences, AIfTN: 1P(,5001 Esnoe~AeadiVriL22233-5

FINAL DISPOSIT(ION: This repor may be destroyed when it is no i--ngci needed. Piw•a c do notreturn it to the U.S. Army Research Institute for the Behavioral and Sociaj Sciences.

NO1 ": Thc findings in this repo,1t arc not to be construed as an officia! I)epartnicut of th-e Arnv1positi, n, urless so designated kv other authorized dmumnlct.s.

For Aiiroe

REPORT DOCUMENTATION PAGE FomB Appo.ved-O

id REPCRr ýElum-;pcc Zs uz - cc. ; S 7iVE MARK.NGS

2a. SECUiRI T C-ASS.; CAC C. C.' -~ 3 DJ.BONi. AVAIL-3dL*7y OF REý,(QR.r

A:,..)rovcd for Dub':.coi..,se;2b O:ECLASS~i.A7T ON/C0UWi%-:ACNC(- SC-ý:Dc.ECsrbuif sU11itU

4 PERFIORM",CZ OR(:AN.ZA7(;N RE?:C4 \(S 5. VONYiOR:NC O,~ Z RE'PORT VVRE;(S)

ARI ecac Rtc

6a- NAME OF PER 0RMiNG 0RGA:.:ZA-C. 60 Oý;;Cp_ SrYMBOL 7a %AVE CF '.OO:CCRGANIZA ON

-\ r--. arhIntt.t (if appiicable)

6c.ADRESS ýCiy, Stare, andc Z,P Code; 7.AZDREP55 Ciry. State. Jr~U Z:P Coc~e)

8a. NA~MF CF 'NG5;cNCRN50"C~ ~O- SY'.%*8L 9 ;ROC'_REVE- ,NSTRUVE'.- ODE%-li:C,%TC% \AB.ORQAN1ZA-ON ~.i&'

1-s't'~ - 3ea. 1:,- 1: i appi~c~le)S.. I s L t

3c. ADDRSS Cay, State, anl Z;PCcje) .0 SOURCE C-- ..NiN V9E.RS

;)CR-. PROJEC7 7ASK WýORK ~~E .l J ;:M:147% 0 NO I ACCE:SS %0 -

VA -522 63027 -793

'(Jr. uce Sec Lr~ty C... ~ssii:cdt:Cfl,

E S. USNA ,' AU .HUN.~)

Ste ~a r~.j~Jh T. Smlc 2._'_: Ia4: R. and Nigel 'R. (AR-.)'3. YPE Or REPOR- 3.TME COVERED O7 CREPORT ý'Year, VO'nrh. Day.' I 11-K PcE COUN T

16. 1-LM\R .OTA7.'CN

17 SCCA2 7--1%i 71 _1111L. TEn e verse 4 ne-zessarl ard aOenriry by .O)cck flumnOO'

G RO C I_ S a. R - :T MA PINT0 Rea~V 1~tŽ %e~c>

-k~,ptra - onal1 Lcst~f in- ManDowcr n ti na-I I::,Miaa f;!ctors Personncll

:9. ý55RAC- ;kContmnue on reverse if -recessaiv anl alerirvy by biock numoer)

The 5. m'. cs.arci . ~t utu(A:%-,) cSvscter.s Rcosecrcli LaiDior torv. (SR-.) apr.t~~~~~~~e~L ir.' ni t :awi a:: c;r-Žn 1 nreza io lM:rRT inLie L e SZ j. n n

U 0 0 1 2:u1 L rluý sts 2~~r' 7 s Isin T (OT'L o f '.e A(:i iU L Rcmc:-oe "v.n i!. I - r Sc~rrn. I: r. e Qýce 1l -)art

M~ 'a 'r. -E r . -.i n*. l i a i~ a ~ ' Ia c r s- l n d D c C n '.'k N.n' t -'-~~'_nL i ,r: J nnd h epinrI' " ;0- n

i 1' L~ ' ~ 1'liic il l: en r ('' -r1,3 OWr ancia.;si,.

L 0 ):STRI83UT'ON lAlA:LAdiJ_;TY A8STJRAC- 21 AI3STLAC- SE(.:uR!Y CA5:FICALOJNC- '. IN CLA S SF: F i l."M: TS -- SAM-/E AS l'T - Dn.c tjSEl~ VS -L ;-si .02 ______________ _-

22a_ NAME OF 5ýElPCr.S:8L-E r.D.V~!)UA_ 22b CFEP.OE (Unchde Aoea Coae) 22c:. ~CE 5YM8O:

CDO Form 1473, JUN 86 Previo~us Paproit ~are')bsolpre. ijiVy CA¶,SjFICiA12CNOF'i-ISAGF

Research Report 1525

MANPRINT Support of Aquila, theArmy's Remotely Piloted Vehicle:

Lessons Learned

John E. Stewart II, Edwin R. Smootz, and Nigel R. NicholsonU.S. Army Research Institute

Manned Systems GroupJohn F. Hayes, Chief

Systems Research LaboratoryRobin L. Keesee, Director

I.S. Army Research Institute for the Behavioral and Social Sciences5001 Eisenhower Avenue, Alexandria, Virginia 22333-5600

Off ice, Deputy Chief of Staff for PersonnelDepartment of the Army

June 1989

Army Project Nurmber Human Factors In Traininq2Q263007A793 and Operational Effectiveness

.'•,p~ov.d to( p-clic i-:,ease. Ostribution is uniimin;!d

iii

FOREWORD

The Systems Research Laboratory (SRL) of the U.S. ArmyResearch Institute for the Be'iavioral and Social Sciences (ARI)supports the Army's MANPRINT (Manpower and Personnel Integration)initiative in the testing and development of new systems and newMANPRINT methods such as HARDMAN II (Hardware vs. Manpower)analysis.

This report discusses the joint effort of two SRL organiza-tions, the Fort Hood Field Unit and the Manned Systems Group, toidentify and resolve major manpower, personnel, and training(MPT) difficulties that reduced the effectiveness of the Army'sRemotely Piloted Vehicle, Aquila. The effort was undertaken inresponse to requests from Headquarters, Training and DoctrineCommand (TRADOC), and the U.S. Army Field Artillery School(USAFAS) in August 1987. In June 1988, a memorandum and a work-ing paper describing the work reported herein were sent to USAFAS(the original proponent for Aquila), the program manager, theTRADOC systems manager, and the U.S. Army Intelligence Center andSchool (the proponent for the Unmanned Aerial Vehicle (UAV)program)

ml-----£o• •:. ... !... lessons~ learned. fr m .. .....

TI JiL ep)Ui L -L L~~ t L- DZV 1~~ II ±Z a A. II ¼:A L 1.£ lt I I L 41 ' i ~.- rr "1111 1

intervention that may be applied to -he successor of Aquila, theUAY. By applying lessons in this way, key MPT issues and con-cerns can be addressed early in the acquisition process and per-formance cf future systems enhanced.

The results of this project were presented at the 30th An-nual Conference of the Military Testing Association in November1988. Although Aquila was cancelled in early 1988, the lessonslearned from this ARI effort were incorporated into the System1-ANPRINT Management Plan for the UAV.

EDGAR. JOHN-SONTechnical Director

V

MWN:PIT 1E ECE A1AUI\, THE AT, ; 'S 1REMOTPEL1Y P11 LOTEI) VFE{1CL-'T-~zO N,;1:A iP

EXECTIVE SUMP1ARY__________ ___ _______

Rcq u -r er.e n :

M-any manpower, personnel, and training (MPT) problems werediscovered during test-ing and. development of Aquila, c.he Army'sRemotely Piloted Vehicle (RPVI). A U.S. Army Research Tns;titutefor the Behavioral and Social Sciences (ARI) Systems Reseaic'La~toratory (SRL) manpower and personnel integration (M-ANPRINT)Task Force, coorainated by the Port Hood Field Unit, was ;-edby H~eadquarters, 'raini~ng and Doctrine Command and the U.S. A-myField A~rtillery schooll to help resolve these YPT problems.

Pr cce6u r e:

1"Th S11- Tas;k- Force conducted an in-depth evaluation of themajor "P1 issues relating to operation and maintenance of theý

RE..Arong ther-e issues were an analysis of specific human --rrd>~ em >covere- c!rn operat jonal liestinln -1 1 (X

Il) , an extcnsýiv, review.. of the literature on imaging systems:and. a sz.7nsiti-itý ana1lysis -applied to the re-sults- -f the original

hor.;revs. rann aw.er (HARDMIAN) analysis to incorporate lessonslearned from 0- !1 -

Findings:

'I ýe SRI, TLask Force was able to identify rajor MPT problem-Sand to suggest concrete ways in which their impact on total sys-ten. performance could be minimized. Most of the recomimendat ionsSuggested MPT sn'utions to the problems.

Utilization of Findings:

Although the Aquila was cancelled in early 1983, the lessonn-learned tro:: this MArNPRIN-T intervention have provided valuable'!insi.ghts intlo the MN'11 issues that must be addressed and res-olvedbefor-e a ne'.. s~y.t~em is- built. Much of this guidance is beingincorooratelJ into the Sytr A~ýN Managjement Plan for the,Unmanned Aerial Vehicle, a family of systems technoLogicallysimilar to Aauila.

v ii

MAN'"RI NT SiP;P.IF O AQVII A, THE ARMY'S RtEMOTELY PtI,OTEI) VEIClEIL:LESSC'N II-ARNED

CONTEN TS

Page

INTRODUCTION . ..................... 1

Background . .MANPRINT . . . . . . . . . . . . . . . . . . . . . . . .

OPERATOR-RELATED MANPRINT CONCERNS ............ ............. 3

Target Acquisition .................. ................... 3Mission Planning . . . . . . . . . . . . . . . . . ..Manpower and Personnel Selection .......... ............ 6

MAINTAINER-RELATED MANPRINT CONCERNS ......... .............

Estinating Maintenance Manpower Requirements .. ....... .Method...................................9

Ae ho . . . . . . . . . . . . . . . . . . . . . . . .

Results and Discussion . . . . . . . . . . . . . . . . . i0Co n cllusions ....................... ............. ...... 12

GENERAL DISCUSSION ................... ..................... 13

Lessons Learned ................... ..................... 13In Retrospect.. .................... ...................... 14

REFERENCES ....................... ......................... 15

LIST CF TABLES

Table 1. Aquila RPV maintenance sensitivity analysis . . . II

LIST OF FIGURES

Figure 1. Aquila battery employment .... ............

ix

111 IPAi1D10- Q A ~EAMY I S RYMOTE:L IA' l 1,'11-DVE1l 1 N : I LESSONS LEARNED

introduction

Bacgqround

Aquila, the Army's Remotely Piloted Vehicle (RPV), is arelatively small unmanned aircraft that is remotely controlledfrom a ground station and is designed to carry a payload that canserve target acquisition, designation, reconnaissance andsurveillance functions. Development of the system can be tracedback to 1971 when the Defense :cience Board recommended that theArmy establish a program which would apply mini-RPV technology tothe fir,, support functions of target designation and adjustmentof artillery fire. The concept of using pilotless aerialvehicles for military purposes was investigated as early as WorldWar I, albeit with little success (Joint Electronic WarfareCenter, 1986). It was looked at again in World War I, and hasbeen under some degree of research and development ever since.Much of the early success was with fairly large target droneswhich simulated flying aircraft and were used for gunneryexercises. Eventually, the idea of modifying such drones byhanging cameras on them emerged, and this combined with evolvingtechnology created many possibilities for military application.For example, the developmenrt of the electronic computer,progressive miniaturization of components, and other relateddevelopments in the 1960s and 1970s, provided the possibilitiesfor building relatively small aii vehicles which couldcontinuously send real time digitized data to a ground station.In addition, it allowed controllers to control actively theflight path or permit the air vehicle to follow a preprogrammedflight path of relatively long duration. Finally, such smallvehicles could be easily launched and recovered, and for somerequirements procured cheaply enough to be expendable.

it was in this technological milieu, combined with anincreasing concentration of Soviet ground-based air defenseweapons posing a high threat to piloted reconnaissance aircraft,that the Defense Science Board nmade its recommendation in 1971 todevelop mini-RPVs for reconnaissance and target acquisition. InSeptember, 1974, the program formally got underway when the U.S.Army Training and Doctrine Command (TRADOC) and the U.S. ArmyMateriel Command signed a letter of agreement to develop an RPVsystem that could demonstrate how RPVs could be used by groundcommanders in recornnais3ance, target acquisition, laserdesignation of targets, and adjustment of artillery fire. Thiseffort evolved into the development of a prototype system whichcame to be known as Aquila and underwent three series of tests:Development Test (DT) II in 1985-1986, (Cozbv, 10986) , OpcrationalTest (OT) II in 1986-1987 [U.S. Army Operational Test andEvaluation Agency (OTEA), 1987], and Force Development Test andExperimentation (FDTE) in late 1987 [U.S. Army TRADOC Test :ncnd

1

Experi ientation Command (TEXCOM), 1988]. Following theconclusion o the FDTE, the Army decided to halt furtherdevelopment and cancelled acquisition plans for Aquila. Tlý'Army's RPV efforts have since been combined with those of theother military services under a Joint Program Office for UnmannedAerial Vehicles (UAV), and the few Aquila systems that wereinitially built for test and evaluation purposes are currentlyserving as a test bed for the joint program.

MANPRINT

Manpower and personnel integzation (MANPRINT) is a new Armyinitiative whose goal is to optimize total system performance byconsidering the soldier as an integral part of the system. Inorder to accomplish this goal, issues and concerns relating tomanpower, personnel, training, human factors engineering, healthhazards and system safety must be addressed early in thedevelopment of a system. In this way the potential adverseimpact of these factors on soldier and system performance can beminimized.

The Army Research Institute (ARI) Systems ResearchLaboratory (SRI) supports the goals of MANPRINT in the testing,development, ai._ acquisiltion of Army systems. This involves notonly support of field testing but also the development and pilot-testing of analytical MANPRINT methods for potentia! users. 1-hecurrent project concerns itself with MANPRINT support to TRADOCprovided by the SRL Aquila MANPRINT Task Force. The ARI FortHood Field Unit was responsible for technical direction of thecurrent project, and had previously provided MANPRINT sipport orOT II. The ART Manned Systems Group was asked to providetechnical assistance in addressing MANPRINT concerns pertinent toAquila's imaging system, and to demonstrate how procedures forestimating maintenance manpower requirements could ':e ihiproved.The ARI Fort Bliss Field Unit provided technical support inaddressing cognitive workload issues.

Lessons learned from this intervention were incorporatedinto recommendations of manpower, personnel and training actionsthat should serve partially to ameliorate some of the problems:hat had threatened 'o diminish the performance of Aquila and

those unmanned aerial systems that will succeed it.

The extensive testing and evaluation that Aquila underwentduring its multiyear development has provided the SRL MANPRINTTask Force with a great deal of useful information about howunmanned aerial systems snould be designed, operated andmaintained. The remainder of this report will focus on some ofthe MANPRINT problems and issues that were brought to lightduring this period. The first section will discuss three topicspertinent to operation of the system: (1) Target Acquisition,(2) Mission) Planning, and (3) Manpower a±nd Personnel Selection.The second will demonstrate how a representative MANPRINT method

2

developed by kPI can be employed as a useful tool for estimatingmaintenance manpower requirements. The results of the cognitiveworkload research will appear in a separate ARI report.

Operator-Related MANPRINT Concerns

Target Acquisition

One of the major findings of OT II was that Aquila crews toooften failed to detect targets. Examination of the results fromOT II (OTEA, 1987) indicates that overall detection rate wasaround 17% (24% of moving targets and 13% of stationary targets).Several uncontrollable extraneous factors contributed to this lowdetection rate. In addition, Aquila crews were required tosearch too large an area during the test. The mean areareconnoitered ducing OT II was 17.6 km 2 , which exceeded the 16km2 area the Army considered to be the maximum for a typicalthree hour Aquila sortie (U.S. Army Field Artillery School,undated). It also became obvious during the test that searchingfor targets from an altitude of 1500 m with a 200 field of viewis very much like looking at the world through a straw, and is avery difficult and tedious process.

Between the completion of OT II and the beginning of the-.... the A. .; rsc-:a_- Aqila sarch -rerm'rPmnntc- uqina data

from OT I! aiid experience learned from other systems and -concluded that it was not feasible to search so large an area fortargets in general. Instead, it seemed better to use otherintelligence sources to indicate approximately where targets maybe, and tc use such information to cue Aquila crews to search asmaller area.

This change was incorporated into the operating proceduresestablished for the FDTE. In addition, software was developed toassist the operator in performing reconnaissance (TEXCOM, 1988).A bookkeeping function was developed which automatically kepttrack of the area that had been searched. During OT II operatorseasily lost track of where they were when they took manualcontrol of the system, thus making it difficult to search an areasystematically. With the new bookkeeping function an operatorcould take manual control of the system, leave the area, performan emergency search somewhere else, and, providing he did notenter a new waypoint, come back and automatically resume thesearch.

The other software improvement was an automatic searchroutine known as the step-stare t,2chnique. This techniquefocused on a given portion of the search area for a fixed amountof time, depending on the amount of clutter in the area, thenautomatically moved to and focused on an adjacent area with abouta 10% overlap, and thus, provided for the systematic search of awhole area. This change increased the time required to search an

3

area, however, from 12 to 35 minutes per km2 . Consequently, theArmy reduced its maximum area search requirement to 6 km2 persortie (U.S. Army Field Art;llery School, 1987).

All of the above changes were incorporated into the Aquilasystem prior to the FDTE with the intention of increasing thetarget detection rate. The results of the FDTE were encouraging.Ninety-eight percent of all moving target airays and 94% of allstationary arrays that were not camouflaged were acquired by theAquila operators during the FDTE (TEXCOM 1988). However, notarget arrays thal were camouflaged were acquired. Thus, theresults showed a definite ability to detect moving andstationary, uncamouflaged targets, and an inability to detecttargets camouflaged with either artificial or natural material.

There were additional targrt acquisition protlems which hadbeen identified during the OT I , and still posed a problemduring the FDTE. For example, operators found it very difficultto detect targets that were in shadows, as often happens when onecamouflages a vehicle by parking it under or near a tree. Dr.Aaron Hyman (1987), recently of ARI, identified a potentialcontributor to this problem, along with a possible solution.Although the camera in Aquila was of very high quality, thecircuitry associated with the automatic gain control and theautomatic level control was designed to clip the upper and lower-a% of th -I-lluI ' -4ancc fzlim) n h ahtcah I 1ir~

transfer function described the relationship between the middle80% of the photocathode illumination and the output signal thatwas transmitted to the ground control station and translated intovarious shades of grey on the video display. When an Aquilaoperator was searching a scene, the camera typically received awide range of input luminances, the lower of which often camefrom dark or shadowy areas. Since the lower 10% of the inputluminances was clipped, the information about objects in theshadows was not transmitted to the video monitors on the ground.

Dr. Hyman suggested a number of ways to solve this problem.A procedural solution would be to use a narrow field of view(FOV), such as 2.70, to search shadowy areas so that the dynamicrange of input luminance spans just the shadowy area. However,this is only feasible for short periods of time. Of the threeAquila FOV options, 2.70, 7.20, and 200, operators tended to usethe 7.20 and 200 FOV's much more than the 2.70 FOV because of thetremendous decrease in the amount of terrain they could observethrough a narrow 2.70 FOV, and the corresponding increase in theamount of time needed to search an area. An engineeringsolution to this problem would be to providý the operator withthe option of switching, as circumstances dictate, to circuitrywhich would not clip the lower input luminances and would employa non-linear transfer characteristic with a high gain at thelower- end of the input luminance distribution. This could beused to amplify arid detect any signals that existed in shadows.

4

Forwdrd Lookinq Infrared (FLIR) uaincras are seen by manyobservers as another solution to this problem of detection oftargets in shadows. However, it should be noted that FLIR, whilehaving many advantages, does not always produce as clear apicture as one would like. When contrast between two objects islow, it is still difficult to distinguish them, and the Armycurrently plans to provide for both daylight TV and FLIR aspotential payloads on its UAV system that is in the initialstages of development. Thus, the problem of detecting targets inshidows with daylight TV cameras is one which must be addressed.

Another aspect of the problem of detecting camouflagedtargets concerns the combination of slant range and FOV forsearching an area. Dr. Hyman's communication with variousexperts in the field of imagery analysis indicated that theacquisition of targets in clutter, such as vegetation, is asomewhat different process than in open terrain. In the lattersituation an observer typically detects something, and then afterobtaining better resolution by moving closer to it or magnifyingit, recognizes it as a particular type of object. In clutter,however, it appears that deteztion of an object or target doesnot occur until resolution is sufficient to recognize it.Previous research (Johnson, 1958) has shown that recognition ofmilitary vehicles with a probability of 0.5 requires about eight(TV) lines of resolution. Thus ii. clutter, it appears that abouteiqht linpq of repolution are needed to detect a tarqet. Dr.Hyman calculated that with Aquila~s 7.20 FOV, a slant ranye of1.66 km was required in order to obtain enemy targetrepresentation (e.g., a tank) of eight lines of resolution.However, Aquila doctrine suggested a 2.5 km slant range for areasearch with the 7.20 FOV. Thus, it is possible that some targetsin clutter were not detected during testing of Aquila becauseadequate target resolution was not obtained at the video monitordue to the non-optimal combination of slant range and FOV.

One way to resolve this problem would be to use a narrowerFOV. This would give more lines of resolution to a target ofgiven size. Another alternative would be to use a shorter slantrancge, such as the 1.66 kin slant range mentioned above, althoughthis would require flying below that altitude which had beendetermined as the minimum for sustained survivability.

igajon Planning

Another problem uncovered during the OT II was that Aquilacrews accepted from higher headquarters mission requests thatexceeded the capabilities of the system (TEXCOM, 1988). Thistypically took the form of crews accepting a mission to search anarea larger than Aquila could reasonably handle during a threehour sortie. The exact cause for this was not empiricallydetermined, but it was thouqht to be due to several factors, toinclude lack of understanding of Aquila's limitations by thecrews, acd a tendency to be unquestioning when receiving mission

5

orders from a higher headquarters. As a result, the trainingprogram was modified prior to the FDTE to give Aquila crewstraining on system limitations and the need to resolve conflictswith higher headquarters cver nonexecutable mission requests.During the FDTE crews were given ten incomplete mission ordersand nine nonexecutable mission orders. They were scored onwhether they resolved those conflicts. Scoring occurred in twoways. First, a subject matter expert (SME), a senior warrantofficer observed the crews during the three phases of a mission(i.e., receiving the mission order at higher headquarters,planning the mission in the Ground Control Station, and executingthe mission by flying the air vehicle and searching for targets)and judged whether or not they resolved mission request conflictswith higher headquarters (TEXCOM, 1988). Second, videotapes ofcrew actions during all three phases were scored by independentevaluators to determine the frequency and distribution of crewactions invclving: a) requests for clarification of missionsfrom higher headquarters, and b) explaining system capabilitiesto higher headquarters (Nicholson, Deignan, & Smootz, 1988).

The results were conclusive. The SME determined that crewsidentified andi adequately resolved all 19 conflicts presented tothem. The videotape analysis showed that this conflictresolution behavior was about evenly spread out over all threephases of mission execution. However, as a proportion of thecommunications activity occurring during each phase, it was foundtha cc-pricd t of the corn i ÷ 'i tion (88%) between Aquilacrewmen and higher headquarters while receiving the mission, butonly 8% of such communication during mission planning, and lessthan 1% during actual flight.

An implication of these findings is that specific trainingin system limitations is extremely important, especially giventhat searching an area with UAVL has proven to be a very slow andtedious process. Individuals in higher headquarters who aretasking 'AVs may easily overestimate their search capability andit is crucial for UAV crews to be trained to expect this and beprepared to resolve such problems with higher headquarters.

Manpower and Personnel Selection

The third topic related to operation of the system focuseson the manpower and personnel requirements for Aquila. There aretwo basic issues here. the rumber of personnel required formanning the system, and the aptitudes required for performing thevarious tasks associated with operating the system.

Information from or II and an independent analysis ofmanpower, personnel and training issues (U.S. Army ResearchInstitute, 1987' indicated that, with respect to the number ofpersonnei requi ed, Aquila was far more manpower intensive th noriginally planned.

6

Operations manpower had its own set of problems. Forexample, the OT II test criterion for preparing and launching anair vehicle after receivinu a mission ordcr was a maximum of 60minutes on 80% of the trials. However, the time criterion wasmet on only 44% of the trials. Part of the failure to reachcriterion can be attributed to the high number of launches thatwere aborted because of maintenance problems indicated by theBuilt-In Test (BIT) system. In fact, throughout OT II anaverage of 2.2 launch attempts were made for every successfullaunch. Nevertheless, the eight man launch and recovery sectionwas usual:y able to handle adequately the launch when equipmentdid not malfunction, and in many cases got a launch off in tenminutes. However, the Or II only tested operations duringdaylight hours. As mentioned earlier, the Army plans to includea FLIR payload on future UAVs and thus provide a 24-hourcontinuous operations capability. Given the high degree ofequipment malfunction that kept crews busy, and the requirementto operate around-the-clock, one is logically led to ask whetheran eight-man crew could continue o function adequately for verylong. Perhaps two four-man crews ;ould, by working 12-hourshifts, but then one must consider that many other duties arisein combat, such as perimeter security. Unfortunately, no datawere collected on this very important manpower problem during anyof the tests, but the SRL Task Force plans to examine it inevaluating future UAV systems.

A riete At-inn that emerged. (and also one on which dataare scarce), concerned the skills and aptitudes re iired foroperating Aquila. A basic problem existed irn that he skillsrequired to operate a syste• like Aquila were distinct enough torequire a unique military occupational specialty (MOS).However, since it was not a high volume system (the Army onlyplanned to acquire nine Aquila batteries requiring about athousand troops), the density of the MOS was low. This situationwas complicated by the fact that soldiers with the Aquila MOSwere assigned to one of two distinct jobs: either operating thelaunch and recovery system, where requirements were ratherphysical, or operating the control station, where requirementswere more cognitive and perceptual. It could be argued that twoMOSs should have been created for these jobs, but doing so wouldhave meant that each MOS would contain only 400-600 troops, avery 'ow density indeed. The Army personnel system hasdifficulty managing low density MOSs. The situation wascompounded by the fact that little empirical data existed on thetypes of aptitudes actually required for operating the controlstation, so it was difficult to make a firm decision as towhether or not that job should require a distinct MOS. Analternative to the creation of a separate MOS is the designationof an additional skill identifier to differentiate thosepersonnel whose job is handling heavy equipment from those whoperform primarily cognitive tasks. These are fundamental issuesto be addressed in futur! UAV MANPRINT efforts.

7

Maintainer-Related MANPRINT Concerns

Estimating Maintenance Manpower Requirements

HARDMAN analysis. In addition to the operations manpowerproblems mentioned in the preceding section, evidence ofpotential maintenance manpower difficulties arose in conjunctionwith DT 1. and OT II. This phase of the Aquila MANPRINT TaskForce eftort was concerned with the application of an analyticalmethod for forecasting manpower requirements, and the impact ofpoor performance of the electronic fault isolation equipmentduring OT II on the accuracy of manpower estimates.

Hardware vs. manpower (HARDY.AN) and its automatedderivative, HARDMAN II are methods for estimating MPTrequirements for emerging systems. HARDMAN was performed on theLockheed Aquila RPV for the Field Artillery School TRADOC SystemsManager under contract with Jet Propulsion Laboratory (DynamicsResearch Corporation, 1983). The original analysis assumed thatthe RPV battery would be fielded in five autonomous sections, andthat the operational scenario would be a 12-hour day.

The 1985 revision of the Target Acquisition, Designationani hAerial __ nnnaiqiance System (TADARS) 'PV Operational andOrganizational (O&O) Plan imparted more cenLralization andcontrol to the Aquila battery. Originally, the battery was toconsist of five autonomous sections, each with the fullcapability of launching, piloting, recovering and maintaining atotal of five air vehicles (AVs); in short, a section couldconduct a complete mission on its own. This arrangement wassuperseded by a more centralized battery, consisting also of fivesections, none of which was fully autonomous. The Aquila batterynow consisted of three Forward Control Sections (FCSs) and tworear area Centralized Launch and Recovery Sections, (CLRSs) eachwith its own Ground Control Station. Each CLRS would have aLaunch Subsystem, Recovery Subsystem, Air Vehicle Handler andfive AVs. It was the responsibility of the two CLR[•s to conductlaunch and recovery operations for the entire battery. After theAV was launched, it was to be handed oft to a FCS, whichcontrolled the aircraft as it carried out its mission. The AVwas to be handed off to a CLRS for recovery upon completion ofits mission. Only one CLRS, the Primary CLRS (or CLRS 1) was tohave a maintenance shelter (MS) along with three spare AVs. TheMS and its crew of four were to provide organizational levelmaintenance for all 13 AVs. Originally, it was intended thateach CLRS have a MS. Figure 1 illustrates the employment of theAquila battery under the CLRS O&O Plan.

8

FCS o-c 11

ICLRS

4-10 10A4C

30LRS4 60A 0

Figure 1. Aqaila battery employment.

Analysis of the original HARDMAN was required to incorporatethese crhanges. One principal finding of the revised HARDMAN(Dynamics Researc1 Corporation, 1985) was that, because of theshift from a 12 te a 24 hour uperat-ional .c.narc. , six -mO IIT Pqmaintainers were required at the MS instead of the four requiredby the 1985 TADARS O&O Plan.

Sensitivity analysis. However, lower than expectedautomatic fault isolation (FI) rates obtained by the automatedtest equipment (ATE) during DT II in 19 6 and OT II in 1987indicated tiat maintenance manpower requirements should bereaddr, ;sed. Also, it appeared that single point estimates ofATE peiformance were not as useful as a ranrge of estimates basedon expected levels of ATE performance.

In an investigation of another Army system, Stewart andShvern (1988) applied HARDMAN II sensitivity analysis to twocomponents of the Forward Area Air Defense system. Maintenancemanpower estimates as a function of automatic fault isolationperformance caused the Army to reexamine its original maintenancemnnpower requirements. Similarly, the Aquila sensitivityanalysis was performed in order to provide the Army with updatedinformation about maintenance requirements for the Aquila RPV,and to illustrate further its efficacy as an adjunct to HARDMANand its derivatives.

Documentation. The principal datA sources were the 1983 and1985 HARDMAN aT,alyses. The 1985 TADA S O&O Plan (with changes)

9

provided information on projected repair times, usage rates, andoperational scenarios. Test data were available from DT II andOT 1I which provided information on maintenance ratios, thenumber of repair actions at organizational and int• rmediatelevels, and operational availability estimates frca these tests.

The Required Operational Capabilities (POC) document calledfor a successful fault isolation (FI) rate of 90%. Generally,ATE performance to date has been poorer than this (see Nauta,1985). During Development Test II (Cozby, 1986), the Aquila ATEsystem only isolated 35% of all faults. During Operational TestII (OTEA, 1987) this rate was slightly less than 20%.

ijnalytical anrproach. The methodology employed in *hepresent analysis was a "top-down" approach which relied onHARDMAN results as a baseline. Becau-.e detailed raw data fromthe HARDMAN were not available, it was ne--essary to rely on datafrom DT II and OT II to obtain estimates of AV down time, repairtimes and i'aintenanca ratios. The wartime operational scenariofrom the 1985 TADARS O&0 Plan allowed for extrapolation to thetotal RPV battery. The resultant annual maintenance manhours(AMMH) obtained through the top-down approach (7722) agreedclosely with those from the HARDMAN (7961 adjusting for the 24hour scenario). Mean Time to Repair (MTTR) times using ATE fromthe revised O&O Plan are wrench-turning times only. Using theAT'. thess wprp: 30 min by day and 45 min by niqht. Manual MTTRswere 90 min (day) and 135 min (night).

The ATE system is mounted inside the MS. For diagnoses offaults to be carried out, the AV must be partially disassembled,defueled, and then moved inside the shelter, accounting for theday-night differences in MTTR.

Operational scenarios. Total AV operating hours for the24 hour-25 mission scenario would be 54 hours, based on the 1935TADARS O&0 Plan.

Maintenance ratios. From operational requirements and OTII test results it can be inferred that 1.21 maintenance actionsper day will be required per air vehicle.

Effects of travel. For the remote CLRS, there should be6.35 maintenance actions per day anticipated. It is assumed thattwo raintainers from the maintenance shelter will retrieve theAV.

Results and Discussion

The sensitivity estimates are presented in Table 1.Dependent variables are annual maintenance manhours (AMMH), finalraintenance ratio (FMR), operational availability (Ao), number ofmiisicn-ready AVs and maintenance man -ears (MMY). Independentvariables are (a) ATE fault isclation (Fl) rates of 90% (similar

10

to the ROC requirement), 40% (slightly better than at DT II) and20% (similar to OT '.I), (b) percentage of repairs performedduring daytime hours ( 50% or 80%) and (c) distance between CLRS1 and 2. Distance between CLRS=0 is equivalent to there beingtwo MSs.

Table 1

Aquila RPV Maintenance Sensitivity Analysis

Distance Between Primary and Secondary CLRS (Km)

0 4 6 8 10 12 14

ATE F1 = 90%(50% daytime repairs)

AMii 7961 8579 8888 9197 9506 9815 10124FMR .40 .44 .45 .47 .48 .50 .51AO .60 .56 .54 .53 .52 .50 .49AVs 7.75 7.34 7.13 6.93 6.72 6.52 6.32,I. 3.32 3.57 3.70 3.83 OP: 4•u• 4.22-

(80% daytime repairs)

AMMI• 5998 6464 6697 6938 7188 7447 7715FIIR .30 .33 .34 .35 .36 .38 .39Ao .70 .66 .66 .65 .64 .62 .61AVs 9.10 8.58 8.58 8.45 8.28 8.06 7.93MMY 2.50 2.69 2.69 2.89 2.98 3.10 3.21

ATE Fl z 40%(50% daytime repairs)

AMMH 13988 14605 14915 15224 15533 15824 16151FMR .71 .74 .76 .77 .79 .80 .82Ao .29 .26 .24 .23 .22 .20 .18AVs 3.77 3.38 3.12 2.99 2.86 2.60 2.34MMY 5.82 6.08 6.21 6.34 6.47 6.60 6.72

11

Table 1 (Continued)

Aquila RPV Maintenan - Sensitivity Analysis.

Distance Between Prinary and Secondary. CLRS (Km)

0 4 6 8 10 12 14

ATE FI = 401

(80% daytime repairs)

AMMH 10966 11818 12243 12684 13141 13614 14104FMR .56 .60 .62 .64 .67 .69 .72AO .44 .40 .38 .36 .33 .31 .28AVs 5.72 5.20 4.94 4.68 4.29 4.03 3.64MM*Y 4.57 4.92 5.10 5.29 5.48 5.67 5.88

ATE EFI = 20%(50% daytime repairs)

Ai 16398 17016 17325 17634 1794! 18202 !Jbvi-F*NIR .83 .86 .88 .90 .91 .93 .94AO .17 .14 .12 .10 .09 .07 .06AVs 2.18 1.82 1.56 1.30 1.17 .91 .78MudY 6.83 7.09 7.22 7.35 7.48 7.61 7.73

(80% daytime repairs)

AXIH 12963 13969 14993 15201 15534 16093 16672FMR .66 .71 .77 .76 .79 .82 .85Ao .34 .29 .24 .23 .21 .18 .15AVs 4.42 3.77 3.38 2.99 2.73 2.34 1.95imY 5.40 5.82 6.25 6.33 6.47 6.71 6.95

Conclusions

Maintenance support of the Aquila would nave posed noproblem if the ATE system were to have perfozeed as specified il,the OO Plan. The sensitivity analyses showed that for ATE F1rates of 90%, the maximum mission requirement (of five AVsairborne at any one time) could have been met with thenaintenance manpower resources available. Neither repairscheduling, the number of MSs, nor distance between CLRS wouldhave posed a threat to operational capability.

12

Were FI rates to fall to 40%, the five-AV requirement couldstill be met under optimal conditions (80% daytime repairs andtwo MSs); with only one MS, the CT.RSs must be no farther than 4kn apart in order to meet the requirement.

If FI rates approxim.,ted OT II results (20%), the missionrequirement could not be met, regardless of assets or schedulingof repair times.

It should be noted that estimates for the present projectagree closely with findings of the draft Human Factorsengineering Analysis (Human Engineering Laboratory, 1987) whichfound that during OT II the MS crew of a battery "minus" haddifficulty keeping up with the workload for a single CLRS. Thereport also expressed doubt that the MS crew would be able tosupport an entire Aquila battery. If operational requirements ofthe Aquila are to be met at all, an FI rate of at least 40%,which is slightly higher thin that attained at PT II, must beachieved, along with the acquisition of an additional MS for eachbattery.

The present analysis further underscores the usefulness ofsensitivity ana]ysis, which enhances the effectiveness of HARDMANII. The uncertain performance of ATE and other electronic faultdiagnostics make single point estimates, based on optimistic

General Discussion

Lessons Learned

Succinctly, it can be stated that the Aquila MANPRINT TaskForce's intervention came too late to save a foundering systemfrom cancellation. In spite of this, major MANPRINT problemsrelating to the operation and maintenance of Aquila werepinpointed. painstakingly analyzed, and concrete recommendationswere made which should provide valuable lessons for thoseinvolved in the development of the current family of UAVs. Theserecommendations illustrated that some MPT problems can bealleviated or at least minimized without requiring expensiveredesign of the system. In fact, with the exception of theimaging system it itself, the difficulties encountered appearedto be attributable more to MPT than to inherent hardware designproblems.

In addition, the effectiveness of sensitivity analysis as auseful adjunct to HARDMAN and HARDMAN II was demonstrated. Oneimportant lesson that carried over from applications of thetechnique to other systems (Stewart & qhvern, 1988) was thatperformance specifications fcr BIT and ATE equipment, at leastwith regard to many new and recently-fielded Army systems, may betoo rigourous. Consequently, the results of HARDMAN IX analyses

13

that assume high levels of BIT and ATE performance become suspectwhen DT and OT results show much lower levels. This is a cogentargument for providing a range of manpower estimates based upon aband of BIT or ATE isolation rates. In such a way decision-makers can determine the minim Ily acceptable level ofperformance that would allow the system to be field within thestipulated manpower constraints that are now imposed by TRADOC.

In Retrospect

In brief, what the ARI MANPRINT Task Force has learned fromits experience with Aquila is izself a good argument forMANPRINT. The Aquila had a protracted developmental historyduring which many MPT and some system design problems wereencountered. One major requirement for successful MANPRINTintervention is timeliness, and it appears that for Aquila, theMPT problems could and should have been addressed much earlier.Had they been, the fate of Aquila may have been quite different.Still, it is reasonable to suppose that similar problems will beencounteied with the UAV Close system, Aquila's successor, whichis currently under development. Now is the time tL. identify,attack, and resolve these problems.

1

14

References

Cozby, R. S. (1986). Test report: Development Test II of theArmy Agi.ila remotely piloted vehicle (RPV) system (U) (Pub.No. USAEPG-FR-1253). Fort Huachuca, AZ: U.S. Army ElectronicProving Ground. SECRET.

Dynamics Research Corporation (1983). AD-lication of the HARDMANmethodology to the Army remotely piloted vehicle (RPV).Wilmington, MA: Contract No. 956-320. Report R-408U.

Dynamics Research Corporation (1985). Areexamination of supportreauirements of the remotely piloted vehicle (RPV) system.Wilmington, MA: Contract No. 956-843. Report R-477U.

Human Engineering Laboratory (1987). Human factors engineeringanalysis for the remotely piloted vehicle (Draft Report).Aberdeen Proving Ground, MD: U.S. Army Laboratory Command.

Hyman, A. (1987). Initial examination of the imagingsystem/operator interface for the Aquila. Memorandum forDirector, Systems Research Laboratory, U.S. Army ResearchInstitute, AlexandrAa, VA.

Johnson, J. (1958). Analysis of imaqe forming systems. Image

Development Laboratories. Fort Belvoir, VA.

Joint Electronic Warfare Center (1986). Unmanned vehiclesyptems renoting technologv study (U) (JON 1-86). SanAntonio, TX: Systems Engirneering Directorate. SECRET.

Nauta, F. (1985). Alleviating fleet maintenance problems throghmaintenance training and aiding research. (NTEC TechnicalReport MDA 903-81-C-0166-1, AD A 155919).

Nicholson, N., Deignan, G., and Smootz, E. (1988). Remotelypiloted vehicle (Acruil r e Development Test andyjuation: Army Research Institute Fort Hood Field Unit

Evaluation (WP FH 88-04). U.S. Army Research Institute FieldUnit, Fort Hood, TX.

Operational Test and Evaluation Agency (1987). IndependentEvaluation of the Remotely Piloted Vehicle (U). IER-OT-604.(Secret; NOFORN)

Stewart, J.E. and Shverri, U. (1988). Apl-ication of HARDMAN I1methodoluqy to the Army's Forward Area Air Defense FAAjsystem. Proceedings of the Human Factors Society 32ndAnnual Meeting, 1117-1121.

15

U.S. Atu1:y Field Artillery S-chool (undated). RemoteKly pilotedv.elii-]e. (Aquila.)_.Operuit ion (Draft FieId Circular 6-MPV)Fort Sil , (K.

U.S. Army Field Artillery School (1987). Remotelypilotedvehicle _Agu~ilaopperation (Draft Field Circular 6-RPV).Fort Sill, OK.

U.S. Army Resealch Institute (1987). Evaluation of ManpgowefrPersonnel and Training sec tions of the Human FactorsEngineering Analysis (HFFAl of the remotely piloted vehicle(RI'V). Alexandria, VA.

U.S. Army TRA)OC Test and Experimentation Command (TEXCOM)(988). Reotej__y_pjoted vehicle: Force Development Testand Experimentation (TEXCOM Test Report FD 0186). FortHcod, TX.