毕业设计论文外文翻译英文文献Word文件下载.docx
《毕业设计论文外文翻译英文文献Word文件下载.docx》由会员分享,可在线阅读,更多相关《毕业设计论文外文翻译英文文献Word文件下载.docx(37页珍藏版)》请在冰点文库上搜索。
学 号:
指导教师单位:
控制工程系
姓名:
职称:
教授
2010年6月15日
Team-CenteredPerspectiveforAdaptiveAutomationDesign
LawrenceJ。
Prinzel
LangleyResearch Center,Hampton, Virginia
Abstract
Automationrepresentsaveryactivearea ofhuman factorsresearch.Thejournal, HumanFactors,published aspecial issueonautomation in1985.Since then,hundredsof scientificstudieshavebeenpublished examiningthenature ofautomationanditsinteractionwith human performance. However,despiteadramaticincrease inresearchinvestigatinghumanfactorsissues inaviationautomation,thereremainareas thatneed furtherexploration。
ThisNASATechnicalMemorandumdescribesanew area ofautomationdesignandresearch,called“adaptiveautomation.”It discussestheconceptsandoutlinesthehumanfactors issues associatedwiththenew methodofadaptivefunctionallocation.The primaryfocusisonhuman-centereddesign, andspecificallyonensuringthatadaptiveautomationis fromateam—centeredperspective. Thedocumentshowsthat adaptiveautomationhasmanyhumanfactorsissues common to traditionalautomationdesign.Muchliketheintroductionofother new technologiesandparadigm shifts,adaptiveautomationpresentsanopportunitytoremediatecurrentproblemsbut poses new onesforhuman—automation interactionin aerospaceoperations. Thereviewhereis intendedto communicatethephilosophical perspectiveanddirectionof adaptiveautomationresearchconductedunder theAerospaceOperationsSystems (AOS), Physiologicaland Psychological StressorsandFactors(PPSF)project.
Keywords:
AdaptiveAutomation;
Human-CenteredDesign;
Automation;
Human Factors
Introduction
"
Duringthe1970sand early1980s..。
theconceptofautomating as much aspossiblewasconsideredappropriate.Theexpectedbenefitwasareduction in pilotworkload andincreasedsafety。
。
.Althoughmany ofthesebenefitshavebeen realized,seriousquestionshavearisenandincidents/accidents thathaveoccurredwhichquestiontheunderlying assumptionsthat amaximumavailableautomation isALWAYSappropriateorthatweunderstandhowtodesignautomatedsystems sothattheyarefullycompatible withthecapabilities andlimitationsof thehumansinthesystem."
---- ATA,1989
The AirTransportAssociationofAmerica (ATA)Flight SystemsIntegrationCommittee(1989)madethe above statementinresponsetotheproliferation ofautomationinaviation.Theynotedthattechnologyimprovements,suchasthegroundproximitywarningsystem, have haddramatic benefits;
others,such as theelectroniclibrarysystem,offer marginalbenefitsatbest。
Suchobservationshaveledmanyinthehuman factorscommunity,mostnotablyCharles Billings(1991;
1997)ofNASA,toassert thatautomationshouldbe approachedfroma”human—centereddesign" perspective.
Theperiodfrom 1970tothe presentwasmarkedbyanincrease intheuseofelectronic displayunits(EDUs);
aperiodthat Billings(1997) calls"information” and“managementautomation。
”Theincreaseduseof altitude,heading,power,andnavigationdisplays;
alertingandwarningsystems,such asthetraffic alert and collisionavoidancesystem(TCAS)andgroundproximity warningsystem(GPWS;
E-GPWS;
TAWS);
flightmanagement systems(FMS)and flightguidance(e.g.,autopilots;
autothrottles)have"been accompaniedbycertaincosts,includinganincreasedcognitiveburdenon pilots,newinformationrequirementsthathaverequiredadditionaltraining,andmorecomplex,tightlycoupled,lessobservablesystems"
(Billings,1997)。
As aresult, human factorsresearchinaviationhas focusedontheeffects ofinformationand managementautomation.The issuesofinterest includeover-reliance on automation, "
clumsy"
automation (e。
g。
Wiener, 1989),digitalversusanalog control, skilldegradation,crewcoordination, anddataoverload(e。
g.,Billings,1997)。
Furthermore, researchhasalsobeendirected towardsituationalawareness(mode&
state awareness;
Endsley,1994;
Woods&Sarter,1991)associatedwithcomplexity, coupling,autonomy, andinadequatefeedback。
Finally, humanfactorsresearchhasintroducednewautomation concepts thatwillneedtobe integrated intotheexistingsuiteofaviationautomation.
Clearly, thehumanfactorsissuesofautomationhave significantimplicationsforsafety inaviation. However,whatexactlydo wemeanbyautomation?
The waywe chooseto defineautomationhas considerablemeaningforhowweseethehuman role in modernaerospacesystems。
Thenext sectionconsiderstheconceptofautomation, followedby anexaminationof human factorsissues ofhuman-automationinteractioninaviation。
Next,apotentialremedy totheproblemsraisedisdescribed,called adaptiveautomation.Finally,thehuman-centereddesignphilosophyisdiscussed andproposalsaremadeforhowthephilosophycanbeapplied tothisadvancedformofautomation。
Theperspective isconsideredin terms ofthePhysiological/PsychologicalStressors&Factorsprojectand directionsforresearch on adaptive automation.
AutomationinModernAviation
Definition.Automation refersto”..。
systemsormethods in which manyofthe processesofproductionareautomaticallyperformedorcontrolledbyautonomousmachinesorelectronic devices” (Parsons,1985)。
Automationis atool,orresource,thatthehumanoperatorcan usetoperform sometask thatwouldbe difficultorimpossible withoutmachine aiding(Billings,1997)。
Therefore,automationcanbethoughtof asaprocessofsubstitutingthe activity ofsomedeviceormachine forsomehuman activity;
oritcan bethought of asastateoftechnologicaldevelopment(Parsons,1985).However, some people(e。
g., Woods, 1996)havequestionedwhetherautomation shouldbe viewed asasubstitutionofoneagentforanother (see"
apparentsimplicity,realcomplexity"
below).Nevertheless, thepresenceofautomationhaspervadedalmost everyaspect ofmodernlives。
From thewheeltothemodernjetaircraft,humans havesought toimprovethe qualityoflife。
Wehavebuiltmachinesand systemsthatnotonly makeworkeasier, more efficient, andsafe,butalso giveus moreleisuretime。
Theadventofautomationhasfurtherenabledustoachieve this end.With automation,machinescannowperformmany of the activitiesthat weoncehadtodo。
Ourautomobile transmission willshiftgearsforus. Our airplaneswill flythemselvesforus.Allwe havetodo is turnthe machineon andoff。
Ithas even beensuggestedthatonedaythere maynotbeaneed forustodoeventhat.However, the increase in“cognitive”accidents resulting fromfaulty human—automationinteractionhave ledmanyin thehuman factorscommunitytoconcludethatsuchastatement may bepremature.
AutomationAccidents. Anumber ofaviationaccidentsand incidentshavebeendirectlyattributedtoautomation.Examples ofsuchinaviationmishapsinclude (fromBillings,1997):
DC—10landingincontrolwheelsteering A330accident atToulouse
B-747upsetoverPacific DC-10overrunatJFK,NewYork
B-747uncommandedroll,Nakina,Ont. A320 accident atMulhouse-Habsheim
A320accident atStrasbourg A300accidentat Nagoya
B—757 accidentat Cali, Columbia A320accident atBangalore
A320landing atHongKong B-737wetrunwayoverruns
A320overrunatWarsaw B-757climboutat Manchester
A310 approachatOrly DC—9wind shear atCharlotte
Billings(1997)notesthateachoftheseaccidentshasadifferentetiology,andthathuman factorsinvestigationofcausesshowthemattertobecomplex.However,whatisclearis thatthe percentageofaccident causeshasfundamentallyshiftedfrommachine-caused tohuman-caused(estimationsof60—80%duetohumanerror)etiologies,andtheshiftisattributabletothechangeintypesof automationthathaveevolvedin aviation。
TypesofAutomation
There areanumberofdifferenttypesofautomationand thedescriptions of themvaryconsiderably。
Billings(1997)offersthe followingtypesofautomation:
•Open-Loop MechanicalorElectronicControl.Automationis controlledbygravityorspringmotorsdriving gears and camsthat allowcontinousandrepetitive motion.Positioning, forcing,andtimingweredictatedbythe mechanismandenvironmental factors (e.g., wind).The automationoffactories duringtheIndustrialRevolutionwouldrepresent thistypeofautomation.
• ClassicLinearFeedbackControl. Automationis controlledas afunctionofdifferencesbetweenareferencesettingofdesiredoutputandthe actualoutput.Changes aremadetosystemparameterstore-setthe automationto conformance。
Anexampleofthistypeofautomationwould beflyballgovernoronthesteamengine。
Whatengineerscall conventionalproportional—integral-derivative (PID)controlwouldalsofitin thiscategory ofautomation.
•Optimal Control。
A computer-basedmodelof controlled processes is drivenby the samecontrolinputs as thatusedtocontrol theautomated process。
Themodeloutputisused toproject futurestates andisthususedto determine thenext control input.A”Kalmanfiltering”approachis usedto estimatethe systemstate to determine whatthe bestcontrolinputshouldbe.
• AdaptiveControl.This typeofautomationactually representsanumber of approachestocontrolling automation,but usually standsfor automation thatchangesdynamically inresponsetoa changeinstate.Examplesincludetheuseof”crisp"
and"fuzzy”controllers,neuralnetworks, dynamiccontrol,andmanyothernonlinearmethods.
LevelsofAutomation
Inaddition to“types”ofautomation,wecanalsoconceptualizedifferent“levels”ofautomationcontrol thattheoperator canhave。