[AD-A122 428 REGISTER OF RESEARCH IN PROGRESS5 ON MENTAL WORKLSAD(U) 2..AIR FORCE AEROSPACE MEDICAL RESEARCH LABWRIGHT-PATT ERSON AFS OH T R METZLER ET AL. JUL 82

UNCLASSIFIED AFAMRL-T R 82 -42 F/G 5/9 NLli~lllllllllI'IIIIIIIIIIIIII,IIIIIIIIIIIIIIffllfllfEIIIIIEEIIIIIEhhmmhhhEEEEEEIlflflllllllllllIIEIIIEEEEIIII

1111.0 ~U I

1.25.

MICROCOPY RESOLUTION TEST CHARTNATIONAL BUREAU OF STANOARDS-1963-A

-.-u- U-

UWSUROF SEAO4m p O5ON MENTAL WOMKLOAD

IL& ARMY A viAI rWARCH AND DsvILopsmET commAND

CLARKA. 8INLDCRAiR Fo=C AEROspACx AbrDCAL RsuAWI LABORAiM)Y

JULY 1962

DTICL AMELECTE

DEC 1GI6 t

AM YOMI AEROSAC IIDCA RNBARO IABOMMTOYAERACM 3MW DV=WAIR F03C3 SF3DUONM4

ILb 16 00*'

-r r16-

When .US Giovermnt drawings. specifications, or other data are aned for mydefnitly elaed Governmnt procarement operation. h, oeuetteey1

nor any obligation whatooever, and the fact that the Goverment am have t~us 1in any way supplied the said drawings, specification. or othe data, Is not to beimplication or otherwise. as in any manner licensing the holder or any other peren rconveying any rights or permission to manufacture, use, or sell any patentd = vetkW*way be related thereto.

Please do not request copies of this report from Air Force Aerospace Medical ResrerhLAdditional copies may be purchased from:

National Technical Information Service528.5 Port Royal RoadSpringfield. Virginia 22161

Federal Government agencies and their contractors registered with Defense Technical InfCenter should direct requests for copies of this repqrt to:

Defense Technical Information CenterCameron StationAlexandria, Virginia 22314

TECHNICAL REVIEW AND APPROVAL

AFAML-TR-82-42

This report him heen reviewed by the C ice of Public Affairs (PA) and is releasable, to tstTechnical Information Service (NTIS). At NTIS. it will be availablo to the general publfie,foreign nations.

This technical report has. been reviewed and is approved for publication.

FOR THR COMMANI)ER

(*HARI^.. HAWKS JR.Chief

Hema EnineeingI~iiisionAir Force Aerospace Andical Monearch Laboatoty

2

X,-.

3ECURITY CLASSIFICATION OF THIS 04SE (111eu, /)at. Entered)I " RED INSTRUCTIONS

REPORT DOCUMENTATION PAGE /BEFRE COMPLETIG OM1. RIFFORT NUMSEW1 2. W VTA SC10.F)OT' CATALOG NUMBER

AFAMRL-TR-82 -42 9&Xff _ T4. TITLE (and Subtitle) S. TF>FP REPOAT &&PERIOD COVERED

REGISTER OF RESEARCH IN PROGRESS ON MENTALWORKLOAD Technical Report

S. PERPORMING ORG. REPORT NUMBER

7. AUTHOR(*) S. CONI RACT OR GRANT NUMBER(*)

*Thomas R. MetzlerClark A. Shingledecker

9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PPOGPAM .LEMENT. PROJECT, TASK

Air Force Aerospace Medical Research Laboratory APeA 6 WORK UNIT NUMBERSAerospace Medical Division (AFSC)Wright-Patterson AFB, OH 45433 62202F; 7184-14-07

I. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE

Air Force Aerospace Medical Research Laboratory J U y 1982Aerospace Medical Division (AFSC) 13. NUMBER OF PAGESWright-Patterson AFB, OH 45433 120

14. MONITORING AGENCY NAME & ADDRESS(if different from Controlling Office) IS. SECURITY CLASS. (of this report)

Unclassified15a. DECL ASSI FICATION/DOWNGRADING

SCHEDULE

IS. DISTRIBUTION STATEMENT (of this Report)

Approved for public release; distribution unlimited.

17. DISTRIBUTION STATEMENT (of the abstract entered In Block 20. it different from Report)

IS. SUPPLEMENTARY NOTES

*Mr Metzler is a member of:U.S. Army Aviation Research and Development Comiand (DRDAV-DM)4300 Goodfellow BoulevardSt. Louis, Missouri 63120

IS. KEY WORDS (Continue on reverse aide it neceaaary ad Identity by block numiber)

Operator WorkloadWorkload Research

20. ABSTRACT (Continue on reverse side It necesamy and identify by block nuinber)

)This report documents current research on operator workload. The register wascompiled from responses to a questionnaire data form sent to 76 scientists whoare active in basic and applied workload research. The six sections of theregister include name and key-term indexes, current project descriptions, list-ings of workload laboratories and potential sources of research support, and abibliography of significant publications in the area.

DO , 1473 ,0,I.ON OF I NOV . IS OSOLETESECURITY CLASSIFICATION OF THIS PAGE ( Data E 1wee40

- -- II I /

PREFACE

This report is an international register of current basic and appliedresearch efforts on the mental workload of human operators. The workinvolved in compiling and editing this report was performed in support ofAFSC Project 7184, Man-Machine Integration Technology for the Air Force, bythe Air Force Aerospace Medical Research Laboratory (AFARL), Human Engi-neering Division, Wright-Patterson Air Force Base, Ohio 45433.

i;i

II0

J A

V~~_ .c04 V

TABLE OF CONTENTS

Page

INTRODUCTION 1

Section I LIST OF CONTRIBUTORS 2

Section II KEYTER4 INDEX 5

Section III CURRENT PROJECTS 20

Section IV WORKLOAD LABORATORIES 89

Section V POTENTIAL FUNDING SOURCES 103

Section VI SIGNIFICANT PUBLICATIONS IN MENTAL WORKLOAD 108

ttI

#i

INTRODUCTION

This volume is the second in a periodic series of reports of currentresearch on operator workload. As in any highly active area of scientificinvestigation, progress toward the understanding of mental workload isimpeded by a less than adequate system of professional communication.Lengthy publication delays, the wide variety of journals and technicalreport sources in which workload research appears, and the sheer volume ofdata being produced substantially detract from the crucial flow of informa-tion among scientists and users of workload research. These problems areultimately reflected in a tendency for needless duplication of effort and areduction in the efficiency with which research findings are translated intoequipment design and training solutions.

The Register of Research in Progress on Mental Workload was conceived as onemeans by which commnunication and cooperation among researchers could beimproved. Essentially, the register is intended to act as a directory toworkload scientists and their projects. For a second year, data forms weresent to individuals selected on the basis of known previous activity in thearea. Seventy-six persons currently engaged in workload research contri-buted responses. The information provided by the researchers was compiledalong with the data obtained in the 1980 edition to produce the six sectionscomprising the 1982 register.

The first two sections provide a name index and a list of key terms whichmay be used to facilitate access to more specific information on individualresearch efforts. The third section contains data on each researcher'saffiliation, mailing address, and telephone number. In addition, each entryis accompanied by current project descriptions provided by the respondent.The fourth section is a listing of laboratories engaged in workloadresearch. The final two sections contain general information deemed to beof possible use to both present and future researchers in the area. Thefifth section lists organizations which, according to the respondents, haveinterests in workload research and may be sources of funding for suchefforts. The sixth section contains responses to a request to list thosepublications, reports, and books which have made the most significant con-tributions to the understanding of mental workload.

Present plans call for the register to be updated and published on a semi-annual basis. Comments on the format of the register, potential improve-ments, and suggestions concerning the inclusion of significant researchefforts not represented in this e-ition are welcomed. Please address anycorrespondence to:

Thomas MetzlerU.S. Army Aviation Research and Development CommandDRDAV-D44300 Goodfellow BoulevardSt. Louis, Missouri 63120

SECTION I

LIST OF CONTRIBUTORS

2

WORKLOAD REGISTER DATA BASE

Alluisi, Earl A.Borg, GunnarBrictson, Clyde A.Brown, IvanButterbaugh, Larry C.Carter, Richard J.Colle, Herbert A.Courtrlght, John F.Crabtree, Mark S.Crawford, Billy N.Damos, Diane L.Donnell, Michael L.Dunn, Richard S.Edwards, Richard E.Eggemeier, Thomas F.Ephrath, Arye R.Fabry, JohnFadden, D. N.Fregley, Alfred R.Gabriel, Richard F.Gartner, Walter B.Gerathewohl, Siegfried J.Goldbeck, Robert A.Gomer, Frank E.Gopher, DanielGriffith, DouglasGunningj, David R.Harris, Sr., Randall L.Hart, Sandra G.Hartzell, E. JamesHelm, Wade R.Hopkin, V. DavidJahns, Dieter W.Jex, Henry R.Johannsen, GunnarJohnson, Edgar M.Kantowitz, Barry H.Kennedy, Robert S.Lane, Norman E.Levine, Jerrold M.Levison, William H.Lindholm, ErnestLinton, Paul M4.McCloy, Thomas M.Mertens, Henry W.Molesko, Norman M.Moray, Neville P.Morgan, Jr., Ben B.Mulder, G.

3

Murphy, Miles R.Navon, DavidNorth, Robert A.O'Donnell, Robert D.Parks, Donald L.Pflendler, ClaudiusQualy, JudiRasmussen, JensReid, Gary B.Rohmert, WalterRoscoe, Alan H.Rosenberg, BruceRouse, William B.Sanders, A. F.Sanders, Mark S.Schiflett, Sam G.Senders, John W.Sheridan, Thomas B.Shingledecker, Clark A.Siegel, Arthur I.Soede, MathijsStein, EarlThiessen, Mary S.Tole, John R.Wickens, Christopher D.Wierwille, Walter W.Wilson, Glenn F.

4

SECTION II

KEYTERN INDEX

A fixed list of keytems was provided on the data form. Respondents arelisted under all keyterms which they indicated as being descriptive of theirresearch interests.

5

II

...... - r"

-

Basic Research

Alluisi, Earl A.Borg, GunnarColle, Herbert A.Crabtree, Mark S.Damos, Diane L.Fregley, Alfred R.Gabriel, Richard F.Gomer, Frank E.Gopher, DanielGriffith, DouglasHarris, Sr., Randall L.Hart, Sandra G.Helm, Wade R.Hopkin, V. DavidJex, Henry R.Kantowitz, Barry H.Kennedy, Robert S.Levine, Jerrold M.Levison, William H.Moray, Neville P.Morgan, Jr., Ben B.Mulder, G.Murphy, Miles R.Navon, DavidNorth, Robert A.Rasmussen, JensRohmert, WalterRouse, William B.Sanders, A. F.Senders, John W.Sheridan, Thomas B.Siegel, Arthur I.Soede, MathijsWickens, Christopher D.Wilson, Glenn F.

Applied Research

Alluisi, Earl A.Borg, GunnarBrictson, Clyde A.Brown, IvanButterbaugh, Larry C.Christensen, Julien M.Crabtree, Mark S.Crawford, Billy M.Donnell, Michael L.Dunn, Richard S.Edwards, Richard E.

6

-------------------"II l i ii llllln lllnili-n lillll ll ....... ... . i...

Eggemeler, F. ThomasEphrath, Arye R.Gabriel, Richard F.Gerathewohl, Siegfried J.Goldbeck, Robert A.Gomer, Frank E.Gopher, DanielGunning, David R.Harris, Sr., Randall L.Hart, Sandra G.Helm, Wade R.Hopkin, V. DavidJahns, Dieter W.Jex, Henry R.Johannsen, GunnarJohnson, Edgar M.Kennedy, Robert S.Lane, Norman E.Levine, Jerrold M.Levison, William H.Linton, Paul M.Mertens, Henry W.Molesko, Norman M.Moray, Neville P.Mulder, G.Murphy, Miles R.North, Robert A.O'Donnell, Robert D.Parks, Donald L.Pflendler, ClaudiusQualy, JudiRohmert, WalterRoscoe, Alan H.Sanders, A. F.Sanders, Mark S.Schiflett, Sam G.Sheridan, Thomas B.Shingledecker, Clark A.Siegel, Arthur I.Soede, MathijsThiessen, Mary S.Tole, John R.Wierwille, Walter W.Wilson, Glenn F.

Theory-Model Development

Borg, GunnerBrown, IvanButterbaugh, Larry C.

7

... . - . . . . .. . . . ... i

Colle, Harbei. A.Crabtree, Mark S.Crawford, Billy M.Donnell, Michael L.Edwards, Richard E.Gabriel, Richard F.Gerathewohl, Siegfried J.Gomer, Frank E.Gopher, DanielHarris, Sr., Randall L.Hopkin, V. DavidJahns, Dieter W.Jex, Henry R.Kantowitz, Barry H.Kennedy, Robert S.Lane, Norman E.Levine, Jerrold M.Levison, William H.Linton, Paul M.Moray, Neville P.Mulder, G.Murphy, Miles R.Navon, DavidO'Donnell, Robert D.Parks, Donald L.Qualy, JudiRasmussen, JensRohmert, WalterRouse, William B.Senders, John W.Sheridan, Thomas B.Shingledecker, Clark A.Siegel, Arthur I.Tole, John R.Wickens, Christopher D.

Metric Development

Alluisi, Earl A.Borg, GunnarBrown, IvanButterbaugh, Larry C.Colle, Herbert A.Crabtree, Mark S.Crawford, Billy M.Donnell, Michael L.Dunn, Richard S.Eggemeer, F. ThomasEphrath, Arye R.Fregley, Alfred R.Gabriel, Richard F.Gomer, Frank E.

8

k. , . " ~...."". . .

Gopher, DanielHarris, Sr., Randall L.Jahns, Dieter W.Jex, Henry R.Kennedy, Robert S.Lane, Norman E.Levine, Jerrold M.Levison, William H.Linton, Paul M.Moray, Neville P.Mulder, G.Navon, DavidNorth, Robert A.O'Donnell, Robert D.Parks, Donald L.Qualy, JudiRoscoe, Alan H.Schiflett, Sam G.Sheridan, Thomas B.Shingledecker, Clark A.Siegel, Arthur I.Tole, John R.Thiessen, Mary S.Wickens, Christopher D.Wierwille, Walter W.

Subjective Measures

Borg, GunnarBrown, IvanButterbaugh, Larry C.Colle, Herbert A.Crabtree, Mark S.Donnell, Michael L.Edwards, Richard E.Eggemeter, F. ThomasGabriel, Richard F.Goldbeck, Robert A.Gunning, David R.Hart, Sandra G.Helm, Wade R.Hopkin, V. DavidJahns, Dieter W.Jex, Henry R.Johannsen, GunnarKennedy, Robert S.Levine, Jerrold M.Molesko, Norman M.Moray, Neville P.Murphy, Miles R.O'Donnell, Robert D.Pflendler, Claudius

9

r/

qualy, JudiRasmussen, JensReid, Gary B.Roscoe, Alan H.Rouse, William B.Sanders, Mark S.Schiflett, Sam G.Senders, John W.Sheridan, Thomas B.Shingledecker, Clark A.Siegel, Arthur I.Wierwille, Walter W.

Performance Measures

Alluisi, Earl A.Borg, GunnarBrictson, Clyde A.Brown, IvanButterbaugh, Larry C.Christensen, Julien 14.Colle, Herbert A.Crabtree, Mark S.Crawford, Billy M.Damos, Diane L.Dunn, Richard S.Edwards, Richard E.Eggemeier, F. ThomasEphrath, Arye R.Fregley, Alfred R.Gabriel, Richard F.Gomer, Frank E.Gopher, DanielGunning, David R.Harris, Sr., Randall L.Hart, Sandra G.Hopkin, V. DavidJahns, Dieter W.Jex, Henry R.Johannsen, GunnarJohnson, Edgar M.Kantowitz, Barry H.Kennedy, Robert S.Lane, Norman E.Levine, Jerrold M.Levison, William H.Linton, Paul 14.Mertens, Henry W.Molesko, Norman M.Moray, Neville P.Morgan, Jr., Ben B.Mulder, G.

10

Murphy, Miles R.Navon, DavidO'Donnell, Robert D.Parks, Donald L.Pflendler, ClaudiusQualy, JudiReid, Gary B.Rohmert, WalterSanders, A. F.Sanders, Mark S.Schiflett, Sam G.Senders, John W.Sheridan, Thomas B.Shingledecker, Clark A.Siegel, Arthur I.Soede, MathtjsTole, John R.Wierwille, Walter W.Wickens, Christopher D.

Physiological Measures

Alluisi, Earl A.Borg, GunnarBrictson, Clyde A.Brown, IvanButterbaugh, Larry C.Ephrath, Arye R.Fregley, Alfred R.Gabriel, Richard F.Gomer, Frank E.Gopher, DanielHarris, Sr., Randall L.Jahns, Dieter W.Jex, Henry R.Johannsen, GunnarKennedy, Robert S.Levison, William H.Mertens, Henry W.Moray, Neville P.Mulder, G.Murphy, Miles R.North, Robert A.O'Donnell, Robert D.Pflendler, ClaudiusQualy, JudiReid, Gary B.Rohmert, WalterRoscoe, Alan H.Sanders, A. F.Sanders, Mark S.Soede, Mathljs

11

I

Thiessen, Mary S.Tole, John R.Wierwille, Walter W.Wickens, Christopher D.Wilson, Glenn F.

Primary Task Assessment

Allulsi, Earl A.Brictson, Clyde A.Brown, IvanButterbaugh, Larry C.Christensen, Julien M.Crawford, Billy M.Dunn, Richard S.Edwards, Richard E.Eggemeter, F. ThomasFregley, Alfred R.Gabriel, Richard F.Gomer, Frank E.Gopher, DanielGunning, David R.Harris, Sr., Randall L.Hart, Sandra G.Jahns, Dieter W.Jex, Henry R.Kantowitz, Barry H.Kennedy, Robert S.Levison, William H.Molesko, Norman M.Morgan, Jr., Ben B.Murphy, Miles R.O'Donnell, Robert D.Parks, Donald L.Qualy, JudiRoscoe, Alan H.Sanders, A. F.Sanders, Mark S.Schiflett, Sam G.Sheridan, Thomas B.Soede, MathijsTole, John R.Wierwille, Walter W.Wickens, Christopher D.

Adaptive Secondary Tasks

Brictson, Clyde A.Dunn, Richard S.Ephrath, Arye R.Fregley, Alfred R.Gabriel, Richard F.

12

Gamer, Frank E.Gopher, DanielJex, Henry R.Kantowitz, Barry H.Levine, Jerrold M.Moray, Neville P.North, Robert A.O'Donnell, Robert D.Schiflett, Sam G.Soede, MathijsWierwille, Walter W.

Nonadaptive Secondary Tasks

Brown, IvanDunn, Richard S.Ephrath, Arye R.Fregley, Alfred R.Gomer, Frank E.Gopher, DanielHarris, Sr., Randall L.Hart, Sandra G.Jex, Henry R.Johannsen, GunnarKantowitz, Barry H.Mertens, Henry H.Moray, Neville P.Morgan, Jr., Ben B.O'Donnell, Robert D.Pflendler, ClaudiusSanders, A. F.Schiflett, Sam G.Senders, John W.Shingledecker, Clark A.Wierwille, Walter H.Wickens, Christopher D.

Continuous Secondary Tasks

Crabtree, Mark S.Damos, Diane L.Dunn, Richard S.Fregley, Alfred R.Gomer, Frank E.Gopher, DanielJex, Henry R.Kantowitz, Barry H.Kennedy, Robert S.Levine, Jerrold N.Levison, Willim H.Mertens, Henry W.Moray, Neville P.

13

/

North, Robert A.O'Donnell, Robert D.Schiflett, Sam 6.Senders, John W.Sheridan, Thoas B.Siegel, Arthur 1.Soede, MathijsWierwille, Walter W.Wickens, Christopher D.

Discrete Secondary Tasks

Alluisi, Earl A.

Brown, IvanIButterbaugh, Larry C.Colle, Herbert A.Crabtree, Mark S.Crawford, Billy N.Damos, Diane L.Dunn, Richard S.Ephrath, Arye R.

Gabriel, Richard F.Gomer, Frank E.Gopher, DanielGriffith, DouglasHarris, Sr., Randall L.Hart, Sandra G.Jex, Henry R.Johannsen, GunnarKantowitz, Barry H.Kennedy, Robert S.Levine, Jerrold N.Mertens, Henry W.Moray, Neville P.IMorgan, Jr., Ben B.Mulder, G.North, Robert A.O'Donnell, Robert D.Pflendler, ClaudiusSchtflett, Sam G.Senders, John W.Sheridan, Thomas B.Slngledecker, Clark A.Soede, MathijsTole, John R.Wierwille, Walter W.

Alternative Assessment Techniques

Ephrath, Arye R.

Fregley, Alfred R.

14

Gomer, Frank E.Jex, Henry R.Murphy, Miles R.O'Donnell, Robert D.Parks, Donald L.Sanders, A. F.Schiflett, Sam G.Senders, John W.Shingledecker, Clark A.Wierwille, Walter W.

Analytical Workload Estimation

Borg, GunnarBrictson, Clyde A.Butterbaugh, Larry C.Crabtree, Mark S.Donnell, Michael L.Dunn, Richard S. fEdwards, Richard E.Fregley, Alfred R.Gabriel, Richard F.Gerathewohl, Siegfried J.Gopher, DanielHopkin, V. DavidJahns, Dieter W.Jex, Henry R.Lane, Norman E.Levine, Jerrold M.Levison, William H.Linton, Paul M.Molesko, Norman M.Moray, Neville P.Murphy, Miles R.Navon, DavidO'Donnell, Robert D.Parks, Donald L.Qualy, JudiRohmert, WalterSchiflett, Sam G.Senders, John W.Shingledecker, Clark A.Siegel, Arthur I.Soede, MathijsWickens, Christopher D.

Scaling

Borg, GunnarBrictson, Clyde A.Butterbaugh, Larry C.Colle, Herbert A.

15

Crabtree, Mark S.Donnell, Michael L.Gopher, DanielHelm, Wade R.Hopkin, V. DavidJex. Henry R.Kennedy, Robert S.Lane. Norman E.Linton, Paul N.Molesko, Norman 14.Moray, Neville P.Navon, DavidO'Donnell, Robert D.Parks, Donald L.Reid, Gary B.Sheridan, Thomas B.Shingledecker, Clark A.Siegel, Arthur I.Wickens, Christopher D.

Validation Research

Borg, GunnarBrictson, Clyde A.Butterbaugh, Larry C.Crabtree, Mark S.Crawford, Billy 14.Donnell, Michael L.Dunn, Richard S.Fregley, Alfred R.Gamer, Frank E.Gopher, DanielHarris, Sr., Randall L.Hopkin. V. DavidJahns, Dieter W.Jex, Henry R.Johannsen, GunnarKennedy, Robert S.Lane, Norman E.Linton, Paul 1N.Moray, Neville P.Mulder, G.O'Donnell, Robert 0.Parks, Donald L.Pflendler, ClaudiusQualy, JudiRasuussen, JensReid, Gary B.Romert, WalterSchiflett, Sam G.Wierwille, Walter W.

16

Aircraft Systems

Allulsi, Earl A.Brictson, Clyde A.Butterbaugh, Larry C.Christensen, Julien N.Crabtree, Mark S.Donnell, Michael L.Dunn, Richard S.Edwards, Richard E.Eggemeler, F. ThomasEphrath, Arye R.Fregley, Alfred R.Gabriel, Richard F.Gerathewohl, Siegfried J.Gomer, Frank E.

Gopher, DanielGunning, David R.Harris, Sr., Randall L.Hart, Sandra G.Helm, Wade R.Jahns, Dieter W.

Jex, Henry R.Johannsen, GunnarKennedy, Robert S.Lane, Norman E.Levison, William H.Linton, Paul M.Molesko, Norman M.Moray, Neville P.Murphy, Miles R.North, Robert A.O'Donnell, Robert D.Parks, Donald L.Pflendler, ClaudiusQualy, JudiReid, Gary B.Roscoe, Alan H.Rouse, William B.Schiflett, Sam G.Sheridan, Thomas B.Shingledecker, Clark A.Thiessen, Mary S.Tole, John R.Wierwille, Walter W.

Surface Transport Systems

Borg, GunnarJex, Henry R.johannsen, Gunnar

17

ji.

Kennedy, Robert S.O'Donnell, Robert D.Pflendler, ClaudiusRohmert, WalterWierwille, Walter W.

Naval Systems

Brictson, Clyde A.Christensen, Julien M.Donnell, Michael L.Gabriel, Richard F.Helm, Wade R.Jex, Henry R.Kennedy, Robert S.Lane, Norman E.Linton, Paul M.Mulder, G.Sanders, A. F.Sanders, Mark S.Schiflett, Sam G.Sheridan, Thomas B.Siegel, Arthur I.

Command Control ard Communications Systems

Alluisi, Earl A.Butterbaugh, Larry C.Christensen, Julien M.Crawford, Billy M.Ephrath, Arye R.Fregley, Alfred R.Gabriel, Richard F.Goldbeck, Robert A.Gomer, Frank E.Gopher, DanielHopkin, V. DavidJohnson, Edgar M.Lane, Norman E.Levine, Jerrold M.Linton, Paul M.Moray, Neville P.O'Donnell, Robert D.Parks, Donald L.Rohmert, WalterSheridan, Thomas B.Shingledecker, Clark A.

Industrial Systems

Borg, GunnarBrown, Ivan

18

Christensen, Julien N.Gabriel, Richard F.Gopher, DanielIHopkin, V. DavidMoray, Neville P.North, Robert A.Rasmussen, JensRohmert, WalterSanders, Mark S.

Weapons Systems

Brictson, Clyde A.Butterbaugh, Larry C.Christensen, Julien M.Crabtree, Mark S.Crawford, Billy M.Donnell, Michael L.Fregley, Alfred R.Gabriel, Richard F.Gomer, Frank E.Gopher, DanielJahns, Dieter W.Jex, Henry R.Lane, Norman E.Linton, Paul M.North, Robert A.O'Donnell, Robert D.Parks, Donald L.Qualy, Judi

Schlflett, Sam G.

19

Section III

CURRENT PROJECTS

This section contains an alphabetically arranged list of researchers. Eachentry includes the researcher's primary organizational affiliation,address, and telephone number, and is accompanied by brief descriptions ofone or more current mental workload research activities of the respondent.

20

Alluisi, Earl A.Chief ScientistAFHRL/CCNBrooks AFB, Texas 78235P:.. (512) 536-3605 (AUTOVAN 240-3605)

STRESS AND PERFORMANCE IN THE FLYING TRAINING ENVIRONMENT

The objective is to investigate stress in the flying training environmentwith particular interest directed to stress of simulated flight and sub-sequent stress of airborne events. Experiments have been planned which(1) allow for quantification of stress experienced in various modes andphases of surface attack training, (2) help determine the interaction ofinstructor techniques and student stress, and (3) provide a clearer pictureof the relationship between stress, learning, and performance in surfaceattack training.

Three studies will examine (1) physiological indices of stress in A-10surface attack training, (2) the effect of IP teaching style on studentstress, and (3) the effect of combat experience on stress response to simu-lated A-10 scenarios.

STRESS UTILIZATION/REDUCTION IN FLYING TRAINING

Previous research established the utility of catecholamine excretion as ametric to assess stress in pilot training. The objective of the presentwork unit is to determine patterns of stress response in pilots exposed tovarying levels of workload/stress in both aircraft and simulated flightscenarios.

Four experiments will be conducted, each of which will assess patterns ofstress response as indicated by a selected set of hormones, neurotransmit-ters, and their metabolities. Urinalysis will be employed. These experi-ments address (1) stress response of pilots in simulated high threat envi-ronments, (2) relationship of psychophysiological and biochemical stressresponses, (3) biochemical response profiles to different stressors, and(4) stress in flight line emergencies.

ASSESSMENT OF WORKLOAD AND PREDICTION OFPERFORMANCE BY COMBINED TECHNIQUES

The objective is to develop and implement combined behavioral and psycho-physiological techniques for measuring pilot attention and task load duringflight simulation. Ultimately these combined techniques will be used tooptimally structure flight simulation training programs and equipment.

A series of laboratory studies will be conducted in order to effectivelydevelop and refine the combined psychophysiological and behavioral measures.These studies will use a simplified flight simulation type behavioral taskwith task difficulty varying as a function of primary and secondary taskdifficulty.

21

___.........____

A series of related flight simulation studies at AFHRL will also beconducted in order to transfer the laboratory findings to the flight simu-lation environment.

AIRCREW PERFORMANCE ASSESSMENT SYSTEM FOR THE C-5 SIMULATOR

The objective is to develop and implement an aircrew performance assessmentsystem for the C-5 flight simulator. The performance measurement systemshall provide a means by which valid, reliable, objective/quantitative mis-sion performance data may be obtained for C-5 aircrew members. These per-formance data shall be obtained for all mission essential/critical perform-ance items. The system as specified shall constitute a feasibility demon-stration or experimental prototype model. As such, it shall provide a meansfor exploring alternative performance measurement strategies and determiningthe relationship between measured performance in the simulator and theaircraft.

Borg, GunnarPTI-Unit, Psykologiska InstitutionenUniversity of StockholmBox 5602, Fiskartorpsvagen 15Stockholm, Sweden S-114 86Ph. 08/22 32 20

SUBJECTIVE ASPECTS OF PHYSICAL AND MENTAL LOAD

During the past 20 years, we have been carrying out studis on varioussubjective aspects of workload. Most of the studies have c,4cerned P02-formance of, and perceptual responses to, heavy physical work, althoughsubjective effort and difficulty perceived during mental tasks have alsobeen dealt with. We have been using three different kinds of stress indi-cators or "effort variables": perceptual, performance, and physiological,which compliment one another. To arrive at a full understanding of man atwork, we must study subjectively perceived differences not only over thewhole range of working conditions, but also in relation to preferred inten-sities of effort, adaptation levels, stress conditions, etc., and observehow subjective changes at these levels are related to performance, physio-logical indicators, and measurements of working capacity (Borg, 1971).

22

Brictson, Clyde A.Dunlap and Associates, Inc.Western Division920 Kline Street, Suite 203La Jolla, California 92037Ph. (714) 459-3377

METHODS TO ASSESS PILOT WORKLOAD AND OTHER TEMPORALINDICATORS OF PILOT PERFORMANCE EFFECTIVENESS

A systematic approach to define, measure, and describe how certain pilot-related variables influence carrier landing performance during sustainedoperations is briefly outlined. Previous exploratory research on the inter-relations between psychophysiological variables, pilot experience, and per-formance is described. Pilot work activity, mood, and sleep are identifiedas indicators of a pilot's temporal state of readiness. A field studydesign and techniques to measure and describe temporal readiness duringprolonged flight operations are provided to demonstrate the methodology inan operational environment. Potential applications of the research arediscussed along with the future role of temporal, psychological, and othermoderator variables in estimating pilot flight status.

AIRCREW PERFORMANCE RESEARCH OPPORTUNITIES USINGTHE AIR COMBAT MANEUVERING RANGE (ACMR)

Three years of aircrew performance measurement related to air combateffectiveness using the Navy's Air Combat Maneuvering Range (ACMR) are pre-sented as evidence of ACMR's research potential. Performance assessmentmethods used to evaluate pilot proficiency are described. The aircrewassessment methods have been used to identify squadron performance differ-ences, evaluate competitive exercises, and provide diagnostic training feed-back to operational users. The use of continuously recorded quantitativemeasures from systems such as ACMR should stimulate more aircrew performancefield research ideas. The availability of objective performance criteriapromises to be of substantial benefit to both the operational user and theresearch community in such areas as pilot selection and training, fleetcombat readiness, and pilot workload and stress.

PREDICTION OF PILOT PERFORMANCE: BIOCHEMICAL ANDSLEEP-MOOD CORRELATES UNDER HIGH WORKLOAD CONDITIONS

A preliminary study of pilot performance, blood chemistry, and sleep/moodpatterns during prolonged periods of carrier operations indicated thatchanges in pilot metabolism and subjective moods can be related to varia-tions in pilot landing performance. A statistically significant multiple Rof .81 (n - 26, pF .O1) was obtained with four predictors. In a separatestudy, pilot sleep patterns for a one-week period were related to landingperformance scores. Pilot intersleep intervals were highly correlated withday landing performance and found to be significantly different than thoseof nonflying personnel.

23

MEASURES OF PILOTS' TEMPORAL STATES OF READINESS

This research follows up previous results with more refined data collectionprocedures and instruments and integrates measures of a pilot's temporalstate of readiness (sleep, mood, workload) with other pilot centered meas-ures. Studies of two stress environments were undertaken: (a) carrierdeployment, and (b) carrier landing qualification (CQ). Pilots from threeattack squadrons embarked on the USS Kennedy (CV-67) during a recent Medi-terranean deployment were the subjects for Phase 1 of the research program.Data were collected on daily activities for three major periods of time: abaseline nonflying period; a follow-up normal flying period; and a 72-hourhigh workload period of continuous flight operations. Daily activity dataon pilot workload (flying and administrative) sleeping, eating, exercise,and mood as well as pilot landing performance for the entire cruise werecollected and analyzed. Findings of interest indicate that pilot workloadaveraged 13 hours per day with 57 percent of the time spent on flight-related activities. When flying increased, squadron administrative work wasreduced proportionally. No group sleep pattern disruptions were notedalthough sleep duration rose as flight activity increased. More short sleepepisodes (naps) during flight periods account for this increase. Perform-ance findings were remarkable for their consistent high levels, especiallyin terms of night carrier landing performance.

Brown, IvanApplied Psychology UnitMedical Research Council15 Chaucer RoadCambridge, England CB22EFPh. 0223 355294

SOME EFFECTS OF TASK AND WORKSPACE DESIGNTRENDS ON CORRELATES OF JOB SATISFACTIONAMONG TELEPHONE SWITCHBOARD OPERATORS

Operators of cordless switchboards report dissatisfaction with their job andperform less efficiently than operators of cord boards. To explore possiblecausal factors, 15 operators participated in a psychophysiological study ofboth working conditions. Cord operators were found: (a) to carry a con-sistently heavier workload of concurrent calls; (b) to be dramatically moreresponsive to diurnal fluctuations in call-traffic level, while maintainingmore stable 'caller waiting times* and "effective call connectionsm; (c) toexhibit this advantage without displaying higher levels of physiologicalstress; and (d) to display consistently higher levels of psychological wellbeing. It is concluded that the cord operators' greater job satisfactionderives mainly from the substantial advantages their flexible system pro-vides them in attaining and maintaining a high quality service.

24

Butterbaugh, Larry C.AFWAL/ FlGRWright-Patterson AFB, Ohio 45433Ph. (513) 255-6931

WORKLOAD PROBLEM ASSESSMENT

The Air Force function for this effort is to research operational aircrewworkload and advanced methods for assessing the workload expected in futurevehicles. A problem has occurred due to the neglect in developing assess-ment methods which are sensitive to workloads in advanced control and dis-play systems. The objective is to maintain familiarity with workloadassessment technology, to have total awareness and understanding of currentoperational workload problems, and to study innovative workload assessmentmethods for their applicability to crew systems engineering. This effortwill contribute knowledge needed to fully understand operational workloadproblems such that the crew systems interface areas requiring further studycan be identified. Operational crews from the major commands were inter-viewed in FY80 and workload related flight problems identified. A data baseof operational workload problems was created in order to identify crewsystem problems that should be addressed in the development of futurevehicles. As a result of this work effort, existing workload problems areidentified for existing system operation, allowing for their study and reso-lution prior to their repetition in future systems (AFWAL-TR-80-3011). Suchinsight into current workload problems and the awareness of improved work-load assessment methods will result in better and more cost-effectivemethods of evaluating crew workload as part of the crew systems developmentprocess. This work effort has been completed.

CREW SYSTEM DESIGN AND EVALUATION METHODOLOGY

The Air Force function supported by this work effort is the research anddevelopment of the crew system interface for advanced airborne weaponsystems. As aircraft missions have become more complex, so has the crew-station design. The point has been reached where the crew system interfaceand its optimization can significantly impact the pilot's ability. Thismeans the crewstation must be designed as an integral unit and that avariety of design evaluations are necessary to assume the design is opti-mized. In other words, the total evaluation must address all the issues(e.g., workload, reliability, performance, fatigue, task allocation, etc.),not just one or two. The objective is to establish and maintain an aware-ness of the complexity of total crew system design and evaluation and tobecome familiar with methods that exist throughout the R&D comunity, and tostudy innovative evaluation methods for their applicability to the crewsystem engineering process. This in-house effort will consist of literaturereviews and data base searches in order to document the range of design andevaluation methods, their constraints for application, and their value as adesign or evaluation tool. The crew system engineering process will bedocumented in TR format in FY81. Periodic, regular comunications willoccur with other DoD organizations in the area of crewstation evaluationtechniques. As a result of this work effort, USAF weapon systems will bemore fully evaluated from the position of the crew system interface. By

25

identifying the range of design issues and methods by which to resolve them.

better and more cost-effective evaluations can occur.

TAACE WORKLOAD PREDICTION

This contracted effort applied an analytic method known as the Controls andDisplays Evaluation Model (CODEM) to flight deck avionics improvements whichresulted from AFWAL/FIGR's Tanker Avionics and Aircrew Complement Evaluation(TAACE) Program. Specifically, CODEM was used to evaluate the capability ofthe TAACE improvements to perform the Rendezvous and In-Flight Refuelingsegments of the KC-135 mission. (TAACE data were utilized because the under-lying intent of this contract was to apply CODE to a previously, independ-ently evaluated design for the purpose of validating the CODEN concept andsensitivity. This evaluation will occur independent from this effort.)

TAACE, which supported the USAF KN-135 Modernization Program, sought toestablish the design criteria for the controls and displays of the improvedflight deck avionics in the event the crew of the KN-135 is reduced byeliminating the navigator. The study involved the analysis of the newflight deck hardware designed for the KN-135; the application of Northrop'sCODEM program to generated crew task complexity, or workload, profiles, andthe identification and validation of changes which, according to the pro-files, would further reduce the crew's workload.

The CODEM results indicated that the flight deck of the modernized KC-135aircraft can be operated satisfactorily in the Rendezvous and In-FlightRefueling segment of a tanker mission with only a pilot and a copilot. Thestudy also established that, on the basis of the CODEM analysis, the newflight deck avionics equipment elements did not cause unacceptably high crewworkloads during the mission segment studied. The technical work has beencompleted and a Technical Report is being processed for publication.

Carter, Richard J.USA Research Institute for the

Behavioral and Social SciencesFort Bliss Field UnitFort Bliss, Texas

DEVELOPMENT OF A SHORT-RANGE AIR DEFENSE SIMULATIONFACILITY AND A CREWMAN PERFORMANCE DATA BASE

BACKGROUND

Studies are currently being undertaken to establish air defense command andcontrol system requirements essential to the accomplishment of the airdefense mission and to evaluate equipment performance and doctrinal con-cepts. An important, critical part of the above system requirements isinformation concerning Short-Range Air Defense (SHORAD) weapon system per-sonnel detection and recognition capabilities. This information is, how-ever, difficult to obtain because of the high cost of live aircraft supportand of the nonavailability of foreign aircraft.

26

During the inevlof 1964 through 1976, research aimed at developing adata bank concerning the capabilities of operators of forward area air

defese eapns o dtec, idntiy, nd stiatethedistance of low-flyingaircraft was conducted. The studies had to be conducted, however, in apart-task evaluation environment. That is, each of these critical tasks was

suidin isolation from the other components of the total operationalsequence. Part-task research only was conducted because techniques formeasuring the effectiveness of the total engagement sequence were not eco-nomically feasible. As a result, there is a lack of information concerningan operator's abilities in performing part-task components when they areembedded in whole-task performance requirements.

RESEARCH

The Army Research Institute has initiated a research program oriented towardthe development of a facility which realistically simulates the SHORADengagement environment and the generation of a data bank of informationconcerning operator performance.

Past Research

The effort thus far has consisted of a paper-and-pencil feasibility study ofthe applicability of low-cost simulation for presenting the SHORAD environ-ment. In Task 1, parameters which must be included in the simulation of theSHORAD engagement environment were identifiled. During Task 2, low-costsimulation approaches were evaluated for their relevance for the SHORADenvironment. Three sets of data that will be necessary to collect whentassessing operator performance capabilities were identified in Task 3.Instrumentation to gather data was also detailed.

Present Research

The present research effort will occur over the course of three years.Year 1 research is directed at fabricating a scaled dynamic flying modelaircraft facility and evaluating the validity of the simulation. The orien-tation of the research in Year 2 will be primarily to gather baseline data.Year 3 research will have the main purpose of investigating the effects ofcueing and early warning on detection and recognition.

Colle, Herbert A.Department of PsychologyWright State UniversityDayton, Ohio 45435Ph. (513) 873-2363

DEVELOPMENT OF ADDITIVE WORK(LOAD SCALES

Fundamental Measurement theory is being applied to develop additive scalesof mental workload. A technique has been developed to equate the workloadof different types of tasks without making arbitrary scaling assumptionsabout the relationship between task performance and mental workload. Theadditivity assumption is being tested both in a secondary task framework

27

and in a subjective scaling framework. Currently, a battery of tasks isbeing evaluated.

Courtright, John F., Ph.D.Acting Chief, Crew Performance BranchUSAF School of Aerospace MedicineBrooks AFB, Texas 78235Ph. (512) 536-3811/3465, (AUTOVON 785-3811/3465)

SWH4ARY OF ON-GOING RESEARCH

The mjor focus of attention for both basic research and exploratorydevelopment for the Crew Performance Branch is aimed at developing andapplying means to assess the stress, fatigue, and workload effects on crewperformance in adverse operational conditions and environments. Both in-house and contract activities are involved. The basic research efforts areperformed as a part of the program of the Air Force Office of ScientificResearch. The exploratory development activities are being undertaken inconcert with the Workload and Erognomics Branch of the Air Force AerospaceMedical Research Laboratory, Wright-Patterson AFB, Ohio.

Current efforts Include development of subjective self-report scales ofperceived operator fatigue for aircrew and groundcrew; subjective self-report scales of perceived operator anxiety states; electrophysiologicaltechniques for the assessment of attention lapses associated with changes ofalertness for aircrew; electrocardiographic techniques for in-flight assess-ment of pilot workload; and methods for assessing the field applicability ofvarious promising laboratory methods of assessing workload presently underdevelopment at the Air Force Aerospace Medical Research Laboratory. Basicresearch efforts are presently directed towards examination of the com-ponents of complex decision-making tasks for their individual componentsusceptibility to the effects of stressors associated with varying demandschedules and unusual work/rest schedules; development of metrics based oneye-blink and eye movement which can be used to index lapses in operatorattention and the onset of fatigue likely to effect decision-making andinformation processing of the type required in the operation of militaryaircraft; examination of the potential for time series analysis of electro-encephalogram and electroculogram data for indexing a pilot's ability toperform information processing and decision-making, particularly as it iseffected by repeated sorties within a given day; and examination of thepotential of sequential optimal control models for becoming a means to indexpiloting behavior in the presence of high cognitive processing workloaddemand under conditions of repeated missions in a single day.

28

tI

Crabtree, Mark S.Systems Research Laboratories, Inc.Human Factors Engineering Group2800 Indian Ripple RoadDayton, Ohio 45440Ph. (513) 426-6000

COMMWUNICATIONS WORKLOAD PROJECT

Most workload assessment techniques suffer from a number of shortcomingsthat can be attributed to the lack of validity and reliability data associ-ated with the techniques, possible difficulties In directly measuring mentalevents, lack of operational applicability, the need for additional equipmentin space-limited environments, and a variety of other ills. Even secondarytasks, which are the most widely used methodology for objective measurementof workload, are often intrusive, unrealistic, and poorly accepted bypilots. A means for overcoming these difficulties may be found in a taskthat is imbedded in other activities that are already a part of the opera-tor's normal duties. Radio conmmunications activities appear to be amenableto such application because they embody many properties of a good workloadmeasure. Information theoretical analysis and pilot opinions indicate thatcommunications activity can be broken down into individual tasks with speci-fiable levels of workload. The results of a simulation study indicate thathuman performance on certain communications tasks is indeed sensitive toprimary task difficulty. Future efforts will be directed toward the valida-tion of these results in high fidelity simulations. The ultimate goal ofthe communications workload project Is the development of a standardized setof comnications tasks that can be used as embedded secondary tasks forworkload assessment at all stages of system evolution.

SECONDARY TASK DEVELOPMENT PROJECT

Som secondary tasks may be more sensitive than other secondary tasks to theworkload imposed by specific primary tasks. It is possible that secondarytasks which share the same mental capacity or resource as the primary taskmay be more sensitive than those that tap separate resources. Oni the otherhand, those that tap the same resource may intrude on the primary task.Therefore, this project was directed toward determining the best combina-tions of primary and secondary tasks. Muich of the researach will be empiri-cal. For example, the memory update task was developed as a secondary taskin which the amount and the rate of information presented to subjects couldbe varied. Analysis of performance scores and SWAT (Subjective WorkloadAssessment Technique) ratings indicated that both variables were effectivelymanipulated. Future studies will concentrate on combining this task withother tasks and determining its sensitivity to workload.

29

Crawford, Billy N.Air Force Aerospace Medical Research LaboratoryTechnology Development Branch, Human Engineering DivisionAFAJ4RL (HEC)Wright-Patterson AFB, Ohio 45433Ph. (513) 255-4379

COMMAND CONTROL WORKLOAD

Workload assessment is included in evaluations of the adequacy of humanengineering in operational Air Force Commnand Control Systems. Subjectiveworkload scales are included in crew opinion surveys to supplement directobservation and measurement of primary task performance. Primary task meas-ures include frequency and duration of crew interactions, telephone/radiocommunications, display viewing and adjustment, and checklist use. Indirectmeasures of task loading are derived from counts of the classes of informa-tion input/output identifiled in functions analyses.

Video/audio tape recordings facilitate assessment of workload indicated byovert behavior. Correlational analysis is used to obtain estimates of theamount of variance in judged workload accounted for by various task activi-ties. Results are used as a partial basis for recoummended human engineeringdesign improvements including changes in procedures, workplace layout,performance aids, task allocations, etc.

SYNTHETIC DATA SYSTEM

Industrial and engineering methods have been evaluated to determine feasiblecandidates for adaptation to Air Force crew workload assessment and alloca-tion. It was determined that a combination of time synthesis methods andstandard data systems was most appropriate. However, significant develop-ment is required. Deficiencies of systems currently used for industrialpurposes are: (1) lack of consideration for divided attention effects, and(2) inadequate data on performance time distributions. Data were collectedon motor-perceptual tasks, perceptual -mediation tasks, display reading andcommunication as a first step toward data base development. Results indi-cate that a "Synthetic Data Systemw Is feasible and that it could be com-bined with computer system simulations and computer-aided design to form aviable approach to workload issue resolution both during and after systemdevel opment. Contractual support for this effort was provided byProfessor James Buck, formerly of Purdue University, now at the Universityof Iowa. Documentation is being prepared for this effort.

30

Damos, Diane L.Department of Industrial EngineeringState University of New York at Buffalo342 Bell HallAmherst, New York 14260Ph. (716) 636-2357

DEVELOPMENT AND TRANSFER OF TIMESHARING SKILLS

Performance on two different task combinations was examined for evidencethat timesharing skills are learned with practice and can transfer betweentask combinations. One combination consisted of two discrete informationprocessing tasks, a short-term memory task and a classification task; theother consisted of two identical one-dimensional compensatory trackingtasks. Three groups of 16 subjects were employed in the experiment. Thefirst received dual-task training on both combinations; the second receivedsingle-task training on the discrete-task combination and dual-task trainingon the tracking combination; the third received dual-task training on thetracking combination only. Evidence for distinct timesharing skills wasfound in both combinations using a new technique designed to separateimprovements in timesharing skills from improvements in single-task per-formance. Transfer of timesharing skills also was found. Several finegrained analyses performed on the data from the discrete task combinationand a control theory analysis of the tracking data indicated that skills inparallel processing were learned in each combination and transferred betweenthem.

INDIVIDUAL DIFFERENCES IN DUAL-TASK PERFORM4ANCE

Eleven right-handed males participated in an experiment examining individualdifferences in multiple-task performance. Three task combinations were usedin the study. The first was composed of a memory task and a classificationtask. The second consisted of two identical one-dimensional compensatorytracking tasks. The third was a dichotic listening task. On day 1 of theexperiment, the subjects practiced each task alone. On days 2, 3, and 4,they performed primarily under dual-task conditio 'ns. Periodically, however,dual-task practice was interrupted to reassess single-task performance. Alldual-task data were analyzed first to determine when stability occurred.Each subject's stabilized data from the tracking-tracking and memory-classification combinations then were corrected for the appropriate single-task baseline. Finally, the subjects were grouped according to which ofthree response strategies they used to perform the memory-classificationtask combination. These strategies were a massed strategy (in which thesubject would emit a series of responses to one task before responding tothe other), an alternating response strategy, and a simultaneous responsestrategy. A two-way repeated measures MANOVA conducted on the stabilizedadjusted data indicated both a significant effect of trials and groups.Possible sources of the between-group differences are discussed.

31

RESIDUAL ATTENTION AS A PREDICTOR OF PILOT PERFORMANCE

Sixteen student pilots performed a task combination consisting of a choi";reaction time task at 1, 2, and 3 bits of information and a one-dimensionalcompensatory tracking task. Cross-adaptive logic was used to keep perform-ance on the tracking task constant, casting the between-subject varianceinto the choice reaction time task. Scores on this combination were corre-lated with performance on flight checks administered after 10, 20, and30 hours of flight training. The multiple correlation between performanceson the flight checks and the task combination increased as the studentsprogressed through flight training. The usefulness of residual attention asa predictor of pilot performance is discussed.

A COMPARISON OF SINGLE- AND DUAL-TASK MEASURESTO PREDICT PILOT PERFORM4ANCE

An experiment comparing the predictive validity of single- versus dual-taskmeasures is reported. Fifty-seven males received two trials on each of twoidentical one-dimensional compensatory tracking tasks. The subjects thenattempted to perform the tasks concurrently for 25 trials. Finally, theyperformed each task alone for one trial. The subjects then were given ashort basic flight course consisting of ground instruction and practice in aGAT-2 simulator. After completing the course, the subjects were asked toperform four repetitions of a descent, a descent followed by a stall, and alevel turn. Performance was scored by an instructor and an observer. Per-formance in the simulator then was correlated with performance on eachtracking trial. The predictive validity of the early single-task scoresdecreased with practice while the dual-task validity increased throughoutthe testing session. However, the predictive validity of the late single-task scores was almost as large as that of the late dual-task scores. Pos-sible explanations for the results are given.

Donnell, Michael L.Decisions and Designs, Inc.8400 Westpark Drive, Suite 600McLean, Virginia 22101Ph. (703) 821-2828

MEASURING SYSTEM EFFECTIVE OPERABILITY

A technique for measuring the effective operability of major air systems[The Mission Operability Assessment Technique (MOAT)] has been developed.MOAT combines the Pilot Task Inventory (PTI) approach with Mu~lti-AttributeUtility Theory (HAUT), which is used to hierarchically organize and deter-mine the importance of tasks and Psychological Measurement Theory andScaling techniques, in particular Conjoint Measurement and the Delta Methodof Scaling. System Effective Operability is defined as the weighted sum ofthe Operabilitles, of the individual tasks. The weights reflect the indi-vidual task's criticalities as well as the importances of the various air-craft phases, subphases, and duty levels which are bases for the hierarchi-cal organization of tasks. The operability of an individual task is anadditive combination of workload and technical effectiveness. Measurement

32

theory and scaling have been utilized to show that an additive combinationis appropriate and to determine the underlying scales. When computerized,MOAT allows the rapid identification and determination of the relativemagnitudes of system operability deficits.

A DECISION-ANALYTIC AID FOR AVOIDING COGNITIVE OVERLOAD

Decisions and Designs, Inc. (DOI), has proposed an approach to workloadmanagement that uses a cost/benefit analysis as its underlying principle.The commander or manager of a complex system, such as a submarine or an airtraffic control center, is viewed as having a limited cognitive capacitywhich must be allocated amongst competing demands on his attention. Whenthe total demand exceeds capacity, he must either shed some of his load orrun the risk of a breakdown in his performance. Cognitive load can be shedin a variety of ways such as selective attention, task simplification, orresponsibility delegation. However, each of these techniques is likely toproduce a poorer result than would be expected if the commander's fullattention were devoted to the task. Thus, shedding cognitive load involvesa cost in that it will produce poorer performance and a benefit in that itreduces the demands placed on the principal decision maker. Using a modelof the costs and benefits associated with shedding cognitive load, a com-puter aid could help a decision-maker determine which of his tasks are mostcritical. The aid could help determine how and when to delegate responsi-bility. As necessary, the aid could help allocate a decision-maker's cogni-tive capacity over the tasks that must be performed. In these ways, the aidcould help avoid cognitive overload and enhance an individual's ability tomanage a complex technological system.

Dunn, Richard S.U.S. Army Research and Technology LaboratoryDAVOL-AS207-5 Ames Research Center

* Moffett Field, California 94035Ph. (415) 965-5579

TACTILE DISPLAY APPLICATIONS

The evaluation of a new tactile display concept required workload compari-sons between visual and tactile display devices and between alternate tac-

* tile display designs and modes of operation. Critical tracking and cross-adaptive subcritical tracking tasks developed by Jex et al., have beenrefined for these purposes. These workload oriented display assessmentprocedures are part of a program to develop tactile displays for operationalhelicopter applications.

GENERAL PURPOSE AND SIMULATOR ORIENTED WORKLOAD MEASURES

A long-term research objective is to develop a variety of workload assess-ment techniques for general use in flight simulator research and other RIDprojects bearing on helicopter design and employment by the Army.Approaches under Investigation include secondary task time estimation, vocal

33

measures of stress, primary and secondary task performance in various con-tinuous and discrete tasks, and analytical procedures for task analysis.The goal is to establish a variety of workload measures to support Armyaviation by making tactical helicopter crew workload a more predictable andmanageable parameter of system design.

Edwards, Richard E.Consulting DivisionBoeing Computer Sevices919 South West Grady WayRenton, Washington 98055Ph. (206) 251-3236

WORKLOAD MODELING

A considerable amount of research has been conducted on workload associatedwith maintaining the aircraft state vector (i.e., tracking tasks) and work-load associated with subystem operation (i.e., discrete tasks). Much of theavailable data, however, is not in a form that can be readily incorporatedinto existing computer based workload models. A series of studies is beingconducted to develop a data base that can be used to extend the capabilityof existing mathematical models of human tracking behavior. In thesestudies, workload associated with manual control of an aircraft's pitch androll axes is being examined under a number of different environmental andsecondary task loading conditions.

WORKLOAD ASSESSMENT

A series of part-task and full-mission flight simulations will be conductedas part of the FAA flight deck certification effort for the Boeing 757 and767 airplanes. The simulation effort will involve both fixed-base andmotion-base research simulators. Objective performance measures and subjec-tive measures will be employed in a direct comparison of workload betweencrewstations with traditional electromechanical instruments and crewstatlonswith advanced electronic displays. These comparisons will be made underboth normal and degraded mode operations.

Eggemeer, F. ThomasDepartment of PsychologyWright State UniversityDayton, Ohio 45435Ph. (523) 873-2444/2391

SUBJECTIVE WORKLOAD ASSESSMENT

In many applications of subjective workload measurement techniques, practi-cal constraints make it.-desirable or necessary to delay completion of sub-jective ratings until some time after actual task performance. This raisesa question concerning the possible effects of such delays on the accuracy ofsubjective ratings. On the basis of the memory literature, it can be pre-dicted that some loss of information will occur at relatively short reten-tion intervals, but little data currently exist that address the specific

34

relationship between retention interval and accuracy of subjective ratingsof workload. A series of studies have been planned to examine the retentioninterval-workload rating relationship. At present, the first study isunderway. In this initial study, subjects perform an information processingtask that requires that they update and recall the status of several con-tinually changing categories of information. Subjects complete a subjectiveworkload rating scale either immediately after the task or at one of tworetention intervals. Immediate ratings will be used as the baseline toassess the effects of retention interval on the delayed ratings. Results ofthis first study will be used to structure subsequent efforts.

Ephrath, Arye R.Bell Laboratories & MIT/Center for Space ResearchBell Laboratories6 Corporate PlacePiscataway, New Jersey 08854Ph. (201) 981-7673

INSTRUMENT SCANNING BEHAVIOR AS ANINDICATOR OF PILOT WORKLOAD

This is an investigation of the relationship between an aircraft pilot'svisual scanning of instruments and his level of mental activity during asimulated approach and landing. This study is motivated by the increasingconcern in several areas of man-machine interaction with the effects ofchanges in manual control and monitoring procedures on mental workload.This concern is particularly keen with regard to airline pilots, air trafficcontrollers, power plant operators, and personnel in control of large ocean-going vessels, since the cost of error can be quite high in any of theseman-machine systems.

Visual scanning behavior plays an important role in each of these systems,since the operator will typically be required to monitor a number of instru-ments which display system state variables. In each of the above roles, thehuman acts as a decision maker, a planner, a manual controller, a monitor,and an event detector. His ability to perform these tasks is generallyinfluenced by their nature, number, and temporal arrangement; by his generalphysical and psychological state; and by the occurrence of unusual or rareevents such as mechanical failures, bad weather conditions, etc. Ideally, ahuman operator's job should be designed in such a way as to require anappropriate fraction of the operator's capacity. To accomplish this designobjective, however, the designer must have a method at his disposal of esti-mating the expended capacity under different conditions. While there existsa number of these methods, none is sufficiently benign and noninvasive to beused in the field (for Instance, in an airliner's cockpit In flight). Con-sequently, we have set out to develop an estimator of mental loading, basedon the operator's visual scan pattern.

In the current work, experiments were conducted in a Terminal ConfiguredVehicle (TCV) fixec .base flight simulator at NASA Langley Research Center.Three NASA test pilots were presented with a piloting task, an arithmetictask designed to vary mental loading, and a side task for calibration of the

353. ...... .. ......

mental loading task. The pilot lookpoint was obtained by using a highlymodified Honeywell oculometer system, and the pilot's eye scan of theinstruments was recorded. The piloting task involved flying a curved Micro-wave Landing System (IiLS) approach from a specified waypoint to touchdown.

The mental loading task was chosen so as not to interfere with the visualscanning of the pilot while providing constant loading during the approach.This was accomplished by having the pilots respond verbally to a series ofevenly-spaced, three-number sequences. The pilot was told that he mustrespond to each three-number sequence by saying either "plus" or "minus"according to the following algorithm: first number largest, second numbersmallest ="Plus"; first number smallest, last number largest - "plus";otherwise ="minus." The numbers were recorded at 20 second and 10 secondintervals. These intervals had been determined empirically to vary mentalloading under a similar piloting task.

The workload measuring side task employed two lights, one Mounted above theother, placed just outside the pilot's peripheral view above the instrumentpanel. The lights came on at random intervals between 1 and .'seconds andremained on for I second. The pilot was told to turn the lights off byusing a three-position rocker switch on the control grip (moving the switchup turned the upper light off, down turned the lower light off). This wasdone only when the pilot had time left from performing the primary task offlying the airplane. Thus, the number of correct responses to the lightssaved a measure of the residual capacity of the pilot from which a workloadindex could be calculated.

A computer algorithm has been developed to obtain the first-order, discrete-~state, discrete-transition, Markov model for each pilot's scanning pattern.It is assumed that workload is constant within each of the six approachsegments since the piloting tasks are essentially constant over each seg-ment. This allows comparison of the instrument transition matrices for eachsegment with those obtained under different loading conditions. The rela-

* tionship between visual scanning and workload is given by the change in theelements of these matrices as loading varies. Higher order Markov modelsmay also be used to provide a more accurate description of the processestaking place.

MENTAL WORKLOAD IN DECISION TASKS

The importance of an operator's mental workload in manual control tasks hasbeen increasingly recognized in recent years. As a consequence, a number ofmental workload evaluation methods have been developed to serve as systemdesign tools for tasks in which the human serves as a control element or inwhich he fills a supervisory (monitoring) role. These methods include per-formance measures, identification of f racti onal -attention parameters inanalytic models, secondary tasks (reserve capacity) techniques, subjectivescaling, and measures of correlated physiological variables.

Research is needed to extend the state of the art from the realm of manualcontrol to that of decision making. With the increasingly evident shift ofthe human's role from that of a manual controller to the executive level of

36

a decision maker, an understanding of the relationships between the demandsimposed by decision making tasks and the resulting cost in mental workloadis essential.

Our approach is to construct an appropriate decision task whose level ofcomplexity and difficulty can be controlled in the laboratory. Methods willbe developed of measuring and quantifying the decision maker's mental work-load in an experimental context. In parallel, the Optimal Decision Model(recently developed in our laboratory) will be expanded to account for taskcomplexity and for the effects of the decision maker's workload level. Theworkload evaluation techniques will be of two categories: (1) reservecapacity index, and (2) performance measures, it, which the decision maker'sperformance (rate of incorrect decisions, for instance) co-varies with hismental workload. The independent (i.e., experimental) variables includeimposed decision rate, processing time requirements prior to decison making(i.e., complexity), number of alternatives, etc. The net objective is theconstruction of a canonical workload task that is readily calibrated interms of noninvasive measures.

Fad"en, D. M.Boeing Commercial Airplane CompanyP. 0. Box 3999Seattle, Washington 98124

The workload assessment techniques used in the design, development, andcertification phases of a commercial airplane program address two basicissues--timeliness of crew actions and ease of operation. The crew must beable to complete all necessary tasks with sufficient reserve time to accom-modate unexpected events. Further, the mental and physical effort associ-ated with each task must not place undue stress on the crew. Both objectiveand subjective methods are used to assess the impact of the flight deckdesign in these areas. Analysis, simulation, and flight test, used singlyor in combination, have proven to be effective techniques for addressing avariety of workload problems. Analytic techniques are of particular inter-est to the aircraft manufacturer since hardware suitable for full scalesimulation or flight test is not available until late in the developmentcycle.

Timeline analysis is the basic means of ensuring that the crew will havesufficient time to accomplish all of their tasks. During design and devel-opment, this technique is augmented by subsystems workload assessments whichevaluate panel layouts and procedures, and by task-time-probability analyseswhich permit variations in task timing and crew performance to be examined.All three are comparative methods enabling a new design to be evaluated withrespect to an existing airplane or subsystem well in advance of any hardwarecommiitments. Following this phase, the physical aspects of workload,associated with hand and eye motion, are confirmed by observation usingmockups, flight simulators, and/or flight test.

The multidimensional nature of what is commonly called mental workloadrequires assessment techniques which provide attribute profiles, not a

37

single index. In this regard we have found specialized subjective ques-tionnaires and rating scales to be the most effective mans for addressingmental workload. While subjective assessment methods are very useful, theycannot provide information in advance of hardware availability. We areexperimenting with information-theoretic estimates of cognitive workload asa means to provide a preliminary evaluation of mental task loading. Such anestimate can provide a common measure for alternative display and controldevices, allowing for a tradeoff between them.

Commercial aircraft workload is qualitatively different than that associatedwith high performance military aircraft. Commercial aircraft development isevolutionary--each new aircraft flight deck draws heavily on past exper-ience. The population of pilots flying commercial aircraft changes veryslowly. The fundamental tasks and decisions expected of the flight crewremain largely the same from one vehicle to the next. It follows, there-fore, that new workload assessment techniques will not have a radical impacton new designs. The primary benefits to be expected from new assessmentmethodologies are improved understanding of the basis for particular designdecisions and greater efficiency in the design process.

Fregly, Alfred R.Air Force Office of Scientific ResearchLife Sciences DirectorateAFOSR/NL Building 410Bolling AFBWashington, DC 20022Ph. (202) 767-5024 (AUTOVON 297-5024)

B I OCYBERNETICS/WORKLOAD

Optimum man-machine sharing and control of increasing information processingloads in high data rate environments will be provided by means of new biocy-bernetic techniques. The machine operator's psychophysiological responseswill be displayed to machines noninvasively, on line in real time via anautomated, adaptive feedback loop. Research emphasis will be given tocoding of brain-evoked potentials by electrophysiological and neuriomagneticmeans. The brain wave signatures determined across a broad spectrum ofsensory, cognitive, and motor activities will serve as templates for devel-oping new neurologically-based metrics of human operator performance. Thiswill assist in identification and prediction of workload factors that influ-ence both the design and operation of crewstations developed for criticaltraining and operational missions.

This program is supported by intramural and contractual efforts at the Aero-space Medical Research Laboratory under Col. Robert O'Donnell to improveman-machine system design and under Andrew Junker in control theory analy-sis; and at the Air Force School of Aerospace Medicine under Dr. ohnCourtright in workload evaluation in mltistress enviroments. The AirForce Office of Scientific Research extramural program is managed byDr. Alfred R. Fregly.

38

Gabriel, Richard F.Human Factors EngineeringDouglas Aircraft Co.Code 35-363855 Lakewood BoulevardLong Beach, California 90846Ph. (213) 593-8642

THE USE OF MENTAL PROCESSING TIME FOR THEEVALUATION OF EADI FORMATS

Integrated displays are commonly evaluated by dynamic tests in flightsimulators using secondary task performance measures and subjective reportsas the basis for evaluation. Both of these measures are of questionablevalidity when applied to mental workload quantification because of testartifacts, biases, etc. This project evaluated an alternative approach thathas resulted from several years of mental workload research at the DouglasAircraft Company. In this approach mental workload is defined as the con-scious mental processes involved in work and quantified as the time requiredto perform the mental processes. Three versions of an EADI were evaluatedby means of discrete tests of the time required for subjects to obtainvarious kinds of quantitative and qualitative information from static CRTdisplays. The results indicate that this is an effective means of evalu-ating alternative display formats during early design phases.

MENTAL WORKLOAD QUANTIFICATION IN A CONTINUOUS TASK SITUATION

The purpose of this current project is to investigate objective methods ofquantifying mental workload involved in obtaining information and makingdecisions from integrated displays. It is a follow-on to a previous inves-tigation of the use of mental processing time measures to evaluate inte-grated displays in a discrete task situation. In the present project, sub-jects are given a flight plan to follow using controls and a dynamic CRTdisplay of flight parameters. The mental processing time required to obtaininformation and make decisions will be measured by the method of subtractionand used to quantify the mental workload involved in alternative displayformats.

PUPILLONETRIC MEASURES OF MENTAL WORKLOAD

A recently completed study evaluated the use of pupillometer (Gulf & WesternApplied Science Laboratories Model 1060-SRP) to assess mental workload.Subjects were given three levels of mental arithmetic under two conditionsof manipulated stress in order to isolate purely automatic effects fromcentral processing factors. The results are being analyzed.

PHYSIOLOGICAL INDICES

The efforts of this project are focusing on two major categories of physio-logical measurement. Electroencephalographic (EEG) and electrooculographic(EOG) measures are being researchedand developed as potentially unobtrusiveindices of aircrew mental workload. A brief listing of research topics is

39

4

given below with a description of the future direction for development. Thelaboratory for EEG and EOG recording is just becoming operational, anddirections for future development may change dependent on what measuresprove useful for operational aircraft, Topics currently under studyinclude:

1. Event Related Potential (ERP), P300 and P165. During simulatedflying, ERPs will be elicited by probe stimuli which are auralwarning events.

2. Iospotential Contour Mapping of Scalp Voltage Levels. Employingmultiple electrode arrays, the contour mapping strategy of mentalworkload research attempts to develop concepts and metrics fordescribing focal brain activity.

3. EOG-Vergence Angle Measurement. Measurement of vergence angles ofpilots with EOG techniques may permit an unobtrusive measurementof external vision in an operational aircraft. Under some circum-stances, the percent time available for external vision (eyesoutside the cockpit) may index overall cockpit efficiency.

4. EOG-Eyeblink. Eyeblink levels are being assessed as indices ofworkload during selected aircraft mission segments. Past effortshave investigated various general physiological responses such asheart rate and heart rate variability, pulse pressure, pulsevolume, etc.

Gerathewohl, Siegfried J.Applied Psychonomics, Inc.5208 Albemarle StreetWashington, DC 20016

Presently preparing a chapter on mental workload in a forthcoming textbookon Military Aviation Psychology for the Department of Defense of the FederalRepublic of Germany.

Gomer, Frank E.Engineering Psychology DepartmentMcDonnell Douglas CorporationP. 0. Box 516St. Louis, Missouri 63166Ph. (314) 576-8283

PHYSIOLOGICAL MONITORING AND MANNED SYSTEM PERFORMANCE

Pilot workload considerations are beginning to have a distinct cognitiveemphasis, due to changes that are occurring in display formats and operatingprocedures. Consequently, new techniques for assessing workload must bedeveloped that are more sensitive to fluctuations in attentiveness and inthe capacity to time-share among several system demands for processing,decision making, and action.

40

_______________________

- - -- w

A flight simulation experiment was conducted in which pilots were trained tofollow a commnanded flight profile and maintain airspeed while concurrentlyperforming threat avoidance and target acquisition tasks. Two types ofbrain electrical activity, event-related potentials (ERPs) and the "on-going" electroencephalogrm (EEG), were analyzed in response to direct miani-pulations of difficulty in executing coumnanded flight maneuvers. Because ofthese manipulations, the difficulty of time-sharing was influenced indi-rectly. Both workload level and continued experience with the flight simu-lation tasks affected the magnitude and latency of the P component of theERP, as well as the distribution of the EEG power. Th99 findings supportthe concept of multidimensional processing resources and the distinctionbetween controlled versus automatic modes of processing.

The results of this investigation are encouraging. They suggest that aphysiological method used to evaluate mental workload in laboratory settingsmay be applied successfully, in at least some instances, to situations inwhich individuals perform operationally relevant tasks.

Gopher, DanielDepartment of Industrial Engineering and ManagementTechnion-Israel Institute of TechnologyTechnionHaifa, IsraelPh. (04) 292-032

A GENERAL APPROACH TO THE MODELING OFTIME-SHARING BEHAVIOR

Human behavior under time-sharing conditions has been conceptualized withina framework inspired by concepts and methods from microeconomy, drawing ananalogy between a manufacturer producing one or more products and a personwith limited resources performing one or more tasks (1,2). This approachhas been used to discuss possible interpretations of dual-task performancewherein concepts such as task difficulty, demand, resource efficiency andresource allocation have been clarified and defined. Special attention isdevoted to the interaction between task characteristics, utility considera-tion, and voluntary control of resources. A survey of the literature and ananalysis of existing models of human capacity have led us to outline amultiple resource approach in which the human processing system is assumedto have a number of mechanisms, each having its own capacity.

Our goal to study the joint effects of difficulty parameters and allocationpolicy resulted in the development of special experimental techniques tomanipulate task priorities in concurrent performance. Response surfacedesign has been employed to enable economical experimental design. Resultshave been analyzed in terms of families of operating characteristic func-tions. Each function within a family depicts the tradeoff of performancebetween two concurrently performed tasks for one condition of task diffi-culty, when resources are allocated in different proportions between thetasks. Experimental tests of these ideas have been conducted using two-dimensional pursuit tracking tasks. These experiments lend support to thetheoretical approach and demonstrate the utility of the methodological tech-niques.

41

Within this general framework, our experiments during the past year weredesigned to substantiate saw of the ideas suggested in the earlier work,clarify the notion of multiple resources, and test the advantages and disad-vantages of such an approach as compared with alternative interpretations oftime-sharing interference. Two main studies mtivated by the results of thefirst group of tracking studies were completed. In one experiment, pursuittracking was paired with binary classification tasks. In another experimenttracking was performed concurrently with variants of a data entry task thatrequired the typing of letter codes.

The first experiment was designed to test the possibility that althoughvertical and horizontal tracking may in many instances not compete for allo-cation of common resources, their coordination in Joint performance taxessome other mechanisms. This involvement of a coordinating mechanism may notbe relevant to tracking performance, yet appear in the performance of athird task. To test this hypothesis, we had subjects make binary classifi-cation of digits while tracking. To guard against possible effects ofperipheral vision, digits were presented with a square that served also asthe control symbol or the target symbol for the tracking task. Memory loadwas manipulated between subjects. Four conditions of tracking were used:single axis, dual axis, dual axis with feedback indicators, and dual axiswith feedback indicators and unequal priorities on the two axes. Experi-mental results failed to reveal the existence of a coordinating resource.Similar decrements in binary classification and tracking performance wereobserved under single-axis and dual-axis tracking conditions. The addi -tional requirement to allocate resources between tracking axes in unequalproportions also failed to affect classification performance. We have,thus, concluded that dual-axis tracking is probably not more costly thansingle-axis tracking in terms of any resource which is not used directly bytracking. An interesting outcome was the overall small deficits that wereobserved on both classification and tracking tasks under dual task condi-tions, which suggests that some of the large deficits reported ins the lit-erature are not due to capacity interference, but are caused by structuralfactors, such as peripheral presentations.In a second experiment, two-dimensional pursuit tracking was paired with aletter typing task in which letters appeared on the screen within the movingsquare of the tracking target symbol and had to be canceled by typing thecorrect letter code on a three-key keyboard. The difficulty of this taskwas varied by incrementing memory load (larger sets of letter codes) orincreasing the difficulty of motor control (selection of more complex motorresponse patterns). The main findings of this experiment were:

1. In single task conditions, both difficulty manipulations yieldedsimilar decrements In typing performance. Manipulation of memoryload tended to cause larger decrements than motor controldifficulty.

2. In dual task conditions, typing performance further deterioratedand decrements were monotonically related to task difficulty.

42

3. Difficulty of motor control interacted with the relative priorityof the typing and tracking tasks while memory load produced aconstant additive increment over levels of priorities.

These results confirm our previous argument that the locus of load in manualcontrol tasks resides in the motor control requirements. The results alsoconfirm the prediction that a multiple resource model in which difficultymanipulations tapping a common resource for concurrently performed taskswill interact with priorities, while manipulations tapping a resource rele-vant to one task only will result in additive effects. At present, we haveseveral experiments underway that seek further clarification of the multipleresource notion and its practical implications for the description of time-sharing performance.

TRAINING PROCEDURES TO IMPROVE TIME-SHARING PERFORMANCE

While the study of general issues relevant to the modeling of time-sharingbehavior is continued, our past year's and current research is marked by anincreased involvement in questions related to training and development oftime-sharing skills and the determinants of improvement in time-sharingperformance. Is. there a general time-sharing factor, whether inherent oracquired, with training? Are different strategies developed for the per-formance of tasks when performed singly or in time-sharing conditions? Whatare the consequences of training subjects in conditions in which intertaskpriorities vary dynamically? These are some of the questions addressed inour studies.

An experiment presently underway in our laboratory addresses the last of theabove questions--training with variable priorities. Two types of trainingconditions are contrasted in this experiment. In one type, tasks are per-formed concurrently and no priorities are indicated, or equal priorities areemphasized and feedback indicators are displayed. In the second types, therelative emphasis on tasks is dynamically varied during training and sub-jects are required to allocate their resources in different proportionsamong the tasks. Training with fixed or unspecified priorities may encour-age development of single optimal interweaving strategy or even integrationof the tasks. Training with variable priorities disrupts integration butsensitizes subjects to the efficiency of resources and the consequences oftheir allocation in various proportions to the concurrently performed tasks.The effects of the two training schedules are investigated, employing pur-suit tracking and the letter typing task described earlier. Partial analy-sis of the results show that the group trained with variable prioritiesdemonstrated larger improvement and higher levels of final performance thanall other groups. When transferred to a new condition in which priorityinstructions and feedback indicators were eliminated, the group trainedunder variable priorities was superior to the other in its ability to shiftresources in order to protect constant performance levels under uninformedvariations of task difficulty. The encouraging first results of theseexperiments are currently substantiated, and further experiments are beingconducted.

43

ULTRADIAN RHYTHM4S IN SUSTAINED ATTENTION TASKS

Short-term rhythmic fluctuations in alertness may be of special significancefor the understanding of the behavior of subjects in sustained attentiontasks. Oscillatory variations in the quality or speed of performance (some-times referred to as lapses of attention) of radar observers, radio opera-tors, air traffic controllers, drivers, pilots, etc., are frequentlyobserved and most commonly treated as a random factor or considered toref lect noisy elements of the human processing system. It may be that theshort-term variability in physiological activity and behavior can be linkedtogether and shown to reveal regularity and organization.

Our research investigates the relationship between 90 minute rhythmic cyclesof physiological processes and the performance of skilled psychomotor tasks.One of the interesting findings in this research is that the physiologicalcycle affects the accuracy of the movements, but does not affect theirspeed. Cyclical increases and decreases in the accuracy measure of motormovements were observed with a peak-to-peak interval of 100 minutes. Cycledeviation corresponded but was not synchronized with physiological pro-cesses. No such changes were observed in the speed of movements. Addi-tional data are being collected for a wide battery of performance tasks.The discovery of ultradlan cycles in performance may lead to many applica-tions in industry.

Gunning, David R.ASO! ENECHWright-Patterson AFB, Ohio 45433Ph. (513) 255-4109 (AUTOVON 785-4109)

KC-135 AVIONICS MODERNIZATION PROGRAM

The goal of the program is to design a two-pilot crewstation for the 135which would eliminate the navigator. A series of simulation studies arebeing conducted to identify the avionics and automation required to replacethe navigator. During the studies, workload measurement is a critical prob-lem. Workload in the existing KC-135 has been measured by recording thetime spent on various cockpit tasks d-iring operational flights by a batteryof measures (i.e., performance, task loading, secondary task, and subjectiver.3tings).

Harris, Sr., Randall, L.NASA Langley Research CenterFlight Dynamics & Control DivisionHampton, Virginia 23665Ph. (804) 827-3871

PILOT WORKLOAD AND SCANINGt4 BEHAVIOR

Pilot workload is composed of at least two components: the physical and themental. Physical workload has been measured for years; however, the mentalworkload is more difficult to quantify. There is a possibility that eyescanning data can be used to quantify mental workload. Laboratory tests

44

have been conducted which have attempted to establish such a link betweenmental workload and eye scanning data. In the tests, mental activity wasaltered by utilizing controlled levels of an auditory side task. The pur-pose of this auditory task was to rob time from the primary flying task. Asthe difficulty of the auditory task increased, the pilot's scanning behaviorwas affected. The pilots stared at the primary instrument and looked atthe peripheral instruments less. In addition, their scanning sequences weredisrupted with the increases in auditory task difficulty. This staring andsequence disruption were more pronounced for the less skilled pilots. Theperformance on the flying task was poorer with increases in auditory task.The results of the laboratory tests suggest that it is desirable to use anauditory task to increase the overall mental workload high enough such thatsystem performance is affected, reflecting differences in display quality orproced