295
1 INTRODUCTION
Many marine accidents are caused by errors on the
partofthenavigator.Inparticular,shipcollisionsare
often causedby improper lookout of navigator. The
improper lookout that is pointed out as the major
cause of ship collisions is regarded as an error in
situation awareness of navigators. Grech and
Horberry indicat
e that 71% of navigators’ errors are
SArelatedproblems(Grechetal.2002).
Inordertodecreasethenumberofshipcollisions,
many analytical studies of marine accidents have
been published (Romer & Petersen 2009, Corovic &
Djurovic2013, Akyuz & Celik 2014). One of these
studies focused on the assessment of navigators’ SA
using the Situation Awareness Gl
obal Assessment
Technique (SAGAT) (Endsley 1988, Koester &
Sorensen 2003). In pilot training using a ship
maneuvering simulator, the possibility that trainees’
navigationskillscouldbemeasuredbySAGATwas
indicated(Okazaki&Ohya2012).
Authorshaveconfirmedthefeat
uresofnavigators’
SAwithbridgesimulatorexperimentsandbehavioral
analysis (Nishizaki & Itoh 2015). However, it is
difficulttounderstand theprioritylevelofotherships
by relying solely on behavioral analysis.
Consequently, in previous study, interviews on
prioritylevelsofrecognizingshipsfornavigatorsare
conducted,andnavigators’SAweremea
suredinship
maneuvering simulator experiments using SAGAT.
Astheresults,anewriskcategorydesignatedasthe
“attention area,” which covered ships with high
priority level in navigators’ SA was proposed
(Nishizaki & Takemoto 2016). The category was
createdbasedontworiskcategories(dangerareaand
A Characteristic of a Navigator's Situation Awareness
for Crossing Ships
C.Nishizaki,T.Takemoto&Y.Kunieda
TokyoUniversityofMarineScienceandTechnology,Tokyo,Japan
ABSTRACT:Manyshipcollisionshavebeencausedbyanavigator’serrorinthesituationawareness(SA)of
thenavigator.Incongestedseaareas,navigatorsclassifyshipsonthebasisofdifferentprioritylevels.For safety
measuresagainstshipcollision,itisimperativefornavigatorstorecognizetheshipswithhighprioritylevels.
In previous study, navigators’ SA was mea
sured in a ship maneuvering simulator using the Situation
AwarenessGlobalAssessmentTechnique(SAGAT).Fromtheresultsofthepreviousstudy,weproposedanew
risk category, named as “attentionarea,” that covers ships withhigh priority level inthe SAof navigators.
However,theextentofdatafornavigators’SAwaslimit
ed.Therefore,thepurposeofthisstudyistoconfirm
thevalidityofthecategoryusingadditionaldataofnavigatorsSA.Inthisstudy,thevalidityoftheproposed
categorywasconfirmed,andalimitlinesurroundingshipswithhighprioritylevelswasidentified.Inaddition,
itwas evidenttha
tthe categorywasabletodetect ships with highprioritylevelaroundthetimewhenthe
collisionavoidancewasperformed.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 11
Number 2
June 2017
DOI:10.12716/1001.11.02.12
296
cautionarea)thatproposedbyKiyoshiHara(Hara&
Nagasawaetal.1990,Hara&Nakamura1995).
We employed only three subjects who had
onboardexperienceascaptainsinpreviousstudy,so
theextentofdataaboutSAinnavigatorswaslimited
in the previous study (Nishizaki & Takemoto 2016).
Therefore,thepurposeofthisstudyistoconfirmthe
validityoftheriskcategoryof“attentionarea”using
additionaldataontheSAinnavigators.
Inthecurrentstudy,foursubjectsemployedwere
different from previous studies. They were
interviewedabouttheirinformallookoutusingsame
listofquestions
asinpreviousstudies,andtheSAin
navigators was measured using SAGAT under the
sameexperimentalconditionsasinpreviousstudies.
The“attentionarea”wasdefinedbya functionof
distance and rate of bearing change of other ships.
Because 3 subjects employed in previous study put
emphasis on them
in determining priority level.
Therefore, in current study, through the interview
questions, it was reconfirmed that navigators
emphasized distance and rate of bearing change of
othershipsindeterminingprioritylevel.
Inaddition,itbecameobviousfromthesimulator
experiments with SAGAT in this study, that the
“attentionarea”covered
shipswithhighprioritylevel
innavigators’SAasinthisstudy.As theresults,the
validityofthe“attentionarea”wasconfirmed,andit
became obvious that there was a limit line covered
ships with high priority level. Furthermore, it was
evident that the “attention area” can assist in
detecting
ships with high priority level around the
timewhenthecollisionavoidancewasperformed.
Inthispaper,themethodsemployedinsimulator
experimentswithSAGATareexplainedinSection2.
Subsequently, the experimental conditions and
scenariosaredescribedinSection3.Theresultsofthe
interviews and simulator experiments are
presented
in Section 4. Based on these results, discussions are
explainedinSection5.Weprovidetheconclusionsof
thisstudyarehighlightedinsection6.
2 METHODFORMEASURMENTOF
NAVIGATORS’SITUATIONAWARENESS
SAGATisaneffectivemethodtomeasurenavigators’
SAs in the simulator experiments. The original
SAGATwas
developedandproposedtomeasurethe
SA of aircraft pilots in cockpits, and the procedural
standardsofSAGATwereestablished(Endsley1988).
ThroughtheuseofSAGAT,itispossibletomeasurea
pilot’s situation awareness directly, independent of
their memory of the experiments after these
experimentsareover.
In
order to maintain the continuity of the
simulation,theinterruptiontimeshouldbeselectedto
be as short as possible. However, there are
ambiguities in the oral reports about all ships in a
congestedseaarea,anditisdifficulttoaccountforall
recognizedshipsinashortinterval.
In
order to measure the situation awareness of
ship navigators, we adopted the method where
subjectsfillinrecognizedshipsonaradarchart.This
method was proposed as an evaluation method for
marinepilottrainees(Okazaki&Ohya2012).Weset
interruptiontimeasoneminutetoenabletestsubjects
toanswerbasedonradarwatchinformationbyuseof
theradarchart.A sample report based on the radar
chartisshowninFigure1.
Figure1.ExampleofareportforSAGAT.
3 EXPERIMENTS
3.1 BridgeSimulator
Thepurposeofthesimulatorexperimentsistoobtain
dataonnavigators’SAinkeepingwatchbytheuseof
SAGAT.ExperimentswithSAGATwereconductedin
the ship maneuvering simulator of the National
MaritimeResearchInstitute.Withinthebridgeofthe
simulator, general navigational equipment
were
installed,suchasacompass,pairofbinoculars,radar,
ECDIS, and steering stand. It is possible to record
behaviors of the subject in response to navigational
orders and other ships situations using behavioral
analysissystemsinthesimulator.
3.2 Subjects
Inthisstudy,foursubjectsdifferentfromtheprevious
study and with onboardexperience wereemployed.
The informationof subjects in a previousstudy and
thatinthisstudyareshowninTable1.
Table1. Comparison between the previous and current
studyonsubjectinformation.
_______________________________________________
Experimental
SeasonSubjectIDAppointment
_______________________________________________
Previousstudy Sub.ACaptain
Sub.BCaptain
Sub.CCaptain
Currentstudy Sub.DChiefOfficer
Sub.E2ndOfficer
Sub.F3rdOfficer
Sub.GChiefOfficer
_______________________________________________
297
In this study, the subjects kept watch using the
general navigational equipment (compass, pair of
binocular, radar, ECDIS and steering stand).
Furthermore, they were instructed to keep their
simulated ships on a steady course and speed if
possible. When the subjects felt an imminent risk of
collision,they wereinstructed
to maneuver to avoid
thecollisions.
Before the experiment, to conform with ethical
standards in humanresearch, we requested that the
all subjects fill informed consent forms for human
research, which all subjects accepted and signed to
signifytheirinformedconsent.
3.3 ExperimentalScenarioandMeasurementMethodfor
navigators’SA
Navigators’SAsweremeasuredusingSAGATunder
the same experimental scenario as used in the
previous study (Nishizaki & Takemoto 2016).
Therefore, an open sea was used as thesea area for
the experimental scenario, and there are 24 other
ships with each having various encounter situations
to the simulated ship
of the navigator. The basic
tracksoftheothershipsareshowninFigure2.
Figure2.ExampleofareportforSAGAT.
In the current study, the SA in navigators was
measuredusingSAGATunderthesameexperimental
conditions as in previous studies. Therefore, the
scenario spanned about 30 minutes including
interruption time for SAGAT, and the experimental
scenario was suspended 4 times after 7 minutes, 12
minutes,17minutes,and22minutes.
Inordertoset
aside enough time to collect information about
surrounding ships, the first measurement time was
setat7minutes.Theinterruptiontimewassettobe
oneminute,andsecond tofourthmeasurementswere
conductedevery5minutes.
Intheinterruptiontime,thesituationawarenessof
the
navigatorswasmea s uredbymeansofthereport,
which the subjects filled in based on the displayed
radar chart (Figure 1). Subjects filled in ships they
recognizedinthereport.Inparticular,aftertheyfilled
in all the recognized ships, the priority rankings of
theseshipswerealsofilledinthe
samereport.
3.4 InterviewafterExperiments
Subjects in current study were interviewed about
theirinformallookoutusingsamelistofquestionsas
in previous studies. The list of questions is divided
into two parts. The first part consists of questions
aboutpersonalhistorysuchaslicenseandnumberof
years
onboard.Thesecondpartdealswithquestions
about informal lookout and situation awareness of
other ships. In particular, four items from the
followingwereincludedinthesecondpart.
Maximumnumberofshipssimultaneously
Radarrangesusedinwatchkeeping
Rankorderlevelofinformation
Rankorder
levelaboutencountersituation
Shortly after the simulator experiments, subjects
were asked to provide their responses to the list of
questionsinturn.
4 RESULTS
In this chapter, we show the results combining
subjectsin previousstudy(Sub.A, Sub.B and Sub.C)
with subjects in current study (Sub.D, Sub.E, Sub.F
and
Sub.G).
4.1 ResultsofInterviewResearch
As a resultof the interviewfor Sub.A to Sub.G, the
maximum number of ships simultaneously
recognizedbynavigatorswasdeterminedtobeabout
five(Table2).Additionally,navigatorsselectedradar
ranges depending on the congestion of thesea area.
The radar ranges used by
navigators are shown in
Table3.
Table2.Maximumnumberofshipssimultaneously
recognized.
_______________________________________________
SeaAreaMaximumnumberofships
_______________________________
SubID A B C D E F G Average
_______________________________________________
OpenSea2 5 3 5 8‐ 5 4.67
CoastalSea4 5 5 5 8 4 5 5.14
Bay(heavytrafficarea) 5 5 5 5 8 4 5 5.29
_______________________________________________
Table3.Radarrangesusedinkeepingwatch.
_______________________________________________
SeaAreaRadarRange[NM]
_______________________________________________
OpenSea12or24
CoastalSea6or12
Bay(heavytrafficarea)3or6
_______________________________________________
In a manner similar to the previous study, four
subjects were asked to provide the types of
informationtheyemployedtodeterminethepriority
levelofotherships.Consequently,thesubjectsranked
the importance of nine sets of information. The
response about the rank order level of information
varied among subjects.
Consequently, data for the
rank order level were normalized to reduce
inconsistencies. Figure 3shows the normalized rank
order level of the aforementioned nine pieces of
information.
Inthisfigure,theordinatedenotesthenormalized
rank order level, and the abscissa denotes the nine
sets of information. The higher the
rank order level,
298
themorecriticaltheinformationwastothesubjectsin
deciding the priority level of other ships. As
representedinFigure3,navigatorsfocusedmainlyon
three pieces of information (the rate of bearing
change, the distance, and the type of encounter
situation)todecidetheprioritylevelofother
ships.
0.5
1
1.5
2
2.5
3
3.5
4
4.5
Changerateofbearing
Distance
Encountersituation
DCPA
TCPA
BCR
Sizeofothership
Statesofmaintenance
Cargo
2.5
2.0
1.5
1.0
0.5
0
‐0.5
‐1.0
‐1.5
Average
Figure3.Normalizedrankorderlevelfornineinformation.
Figure4. Normalized rank order level about encounter
situation
Consequently,theprioritylevelofshipsshouldbe
investigatedineachencountersituation.Furthermore,
itisimportanttoanalyzeSAinnavigatorsbased on
therateofbearingchangeanddistance.Theseresults
were comparable with the results of the previous
study.
Inaddition,eachsubjectrankedtheimportanceof
ten
encounter situations. Figure 4 shows normalized
rank order level for ten encounter situations of all
sevensubjects. Inthisstudy,ahigherrankorderlevel
indicatesamorecriticalencountersituationtodecide
theprioritylevelofotherships.
Figure4indicatesthatnavigatorsplacedemphasis
oncrossingsituationsfrom
starboardascomparedto
otherencountersituations.Theresultsobtainedinthe
currentstudycorroboratethatofthepreviousstudy.
4.2 Resultsofnavigators’SAMeasurements
Measurementresultsofnavigators’SAareshownin
Table4.ThefirstcolumninTable4indicatesshipID
includingthescenario.Thesecondand
thirdcolumns
indicate encounter situations of the other ships. The
right side beyond the fifth column indicates the
resulting navigators’ SA. The total number of ships
were recognized by each subject at the time of
measurement is indicated in the 3rd row. As I
mentioned before, the maximum number of
ships
simultaneously recognized by navigators was
determined to be about five. Therefore, we defined
ships ranked higher than 5th as ships with high
prioritylevel.Circlesymbols(o)areusedtoindicate
ships ranked higher than 5th and cross symbols (x)
are used to indicate ships ranked 6th or lower in
Table 4. No symbols (blank space) were used to
indicateshipsthatwentunrecognizedbyanyone.
Inaddition,Table4showsthatthereareshipsthat
arerecognizedbyseveralsubjects.Forexample,Ship
3inthe2ndand3rdmeasurementsisrecognizedby
allsubjects. Ontheother hand,
Ship11 and Ship21
were not recognized by any of the subjects.
Interestingly,despiteship4in4thmeasurementbeing
recognizedbyseveralsubjects,itrankedashavinga
low priority level. From these results, it can be
deducedthatmanyofthesubjectsagreedonsimilar
shipshavinga
highprioritylevel.
5 DISCCUSSIONS
Based on the interviews, it became evident that
subjectsplacedmoreemphasisoncrossingsituations
from starboard as compared to other encounter
situations.Therefore,inthispaper,wefocusedonfive
shipsthatcrossaheadofthesubjects’ownshipsfrom
starboardtoport(ship
IDsforsuchships are3,4,10,11
and16)
Furthermore, based on the observations of the
interviews, it was considered to be important to
analyzetheSAofnavigatorswiththechangingrates
of bearing and distance. Adapting from previous
work reported by Nishizaki (Nishizaki & Takemoto
2016),theemphasisispla ced
ontheratesofbearing
anddistance.
299
Table4.Resultsofnavigators’situationawareness.
__________________________________________________________________________________________________
MeasurementNo.(time) 1(7minutes)2(12minutes)3(17minutes)4(22minutes)
_________________________________________________________________________________________
Ship RelationshipsSubjectID A B C D E F G A B C D E F G A B C D E F G A B C D E F G
________________________________________________________________________________
IDTotalNum.
ofShips 5 6 6 7 5 4 5 8 6 5 6 7 4 4 7 7 7 4 6 3 4 7 6 7 5 3 3 3
__________________________________________________________________________________________________
1 StoP
*1
CrossO
2 S OvertakeO O O O X O X O O O O O O O O O
3 StoP CrossO X O O O O O O O O O O O O O O O O O O O O O
4 StoP CrossO O O O O O O X O O X X
5 PtoS CrossO O O O
6 PtoS CrossOO O O O O OO O
7 PtoS CrossO O
8 PtoS CrossO O O O O O X O O
9 StoP CrossO XO O O OO X O
O
10 StoP CrossO O O O O OO O O O O O O O O
11 StoP Cross
12 PtoS CrossO OO
13 PtoS CrossXX
14 PtoS CrossOO
15 StoP CrossX X
16 StoP CrossXO
17 PtoS CrossO
18 P HeadonXXO
19 S ParallelO OO X OO XO O
20 StoP CrossX X X O O
21 S Headon
22 S ParallelO
23 S HeadonX OO O O O O
24 StoP CrossO O OO O O XO X O OO
__________________________________________________________________________________________________
*1“S”denotesthestarboardsideand“P”denotestheportside
Figure5.Resultsofnavigators’ situationawarenessandattentionarea.
Figure5shows resultsofSAinnavigatorsforfive
shipsthatcrossaheadofthesubjects’ownshipfrom
starboardtoport.InFigure5,thex‐axisindicatesthe
distancetootherships,whilethey‐axisindicatesthe
rate of bearing change. The right side of the figure
indicates the likelihood of the other ship crossing
aheadofthesubject’sshipinthefuture,whiletheleft
side of the figure indicates if this other ship has
alreadycrossedaheadofthesubject’sship.InFigure
5,theblackanddarkgraycoloredareasshowdanger
and caution
areas of Hara’s risk categories,
respectively (Hara & Nagasawa et al. 1990, Hara &
Nakamura 1995). It was observed in the previous
study, that some ships were collectively sighted
/agreed upon by several subjects as having a high
priority level. Thus, a new risk category: the
“attentionarea,”wasproposed.The
attentionareais
representedasthelightgraycoloredareainFigure5.
Black circle symbols were used to indicate high
priority level ships and cross symbols were used to
300
indicate low priority level ships or ships
unrecognizedbyanyone.Inthisstudy,somesubjects
maneuvered to avoid collisions in the simulator
experiments. Therefore, there are variations between
theplotofthecurrentstudyandthatoftheprevious
study(Nishizaki&Takemoto2016).However,results
oftheSAmeasured
atthetimeofavoidanceactions
werenotincludedinTable4andFigure5.
Itisobviousthattheattentionareacoveredships
withhighprioritylevelinSAofnavigatorsincurrent
study.Therefore,the validityof the “attention area”
was confirmed. Furthermore, results indicate that
therewas
alimitlinethatcoveredshipswithahigh
prioritylevel.TheinterruptedlineinFigure5shows
the limit line surrounding high priority level ships.
The corresponding equation and weighing
coefficients of attention area and the limit line are
showninTable5.
istherateofbearingchange,R
is the distance, while,
,
,
and
are
weighingcoefficients.
In this study, some subjects maneuver to avoid
collisions in the simulator experiments. Due to the
abilityofthesubjectstomaneuverusingthesimulator
without any collisions, the assumption that the risk
category of attention area was available to detect
shipswithhighpriority levelaround
the time when
thecollisionavoidancewasperformed.
Table5.Equationandweighingcoefficients.
_______________________________________________
EquationWeighingCoefficients
_______________________________________________
RR
36.18.0
,108.5
,7.1,108.5
5
5
areaattention
_______________________
45.185.0
,108.5
,7.1,108.5
5
5
LineLimit
_______________________________________________
6 CONCLUSION
In order to confirm the validity of the risk category
(attention area), subjects employed were different
from a previous study, and navigators’ SA were
measured with the SAGAT under the same
experimentalconditionsinpreviousstudy.Basedon
the interviews and measurements of the navigators’
SA,the
followingwereconfirmed:
1 Itwasreconfirmedthatnavigatorsprioritizedthe
distanceandrateofbearingchangeofotherships
indeterminingaprioritylevel.
2 Theriskcategory(attentionarea)coversshipswith
high priority level in SA of navigators in this
study. Thus, the validity of the “attention area”
wasconfirmed.
3 Thereisalimitlinethatincludesshipswithhigh
prioritylevel.
The attention area was available to detect ships
witha highprioritylevel around the timewhen the
collisionavoidancewasperformed.
Because size of navigators’ SA data in previous
and current study are very
small, it is necessary to
verify the validity of attention area by increasing
navigators’SAdata.
ACKNOWLEGMENT
This work was supported by JSPS KAKENHI Grant
NumberJP15K16304.
REFERENCES
Akyuz,EandCelik,M.2014.UtilisationofCognitiveMap
inModellingHumanErrorinMarineAccidentAnalysis
andPrevention.SafetyScience.Vol.70:19‐28.
Corovic, B. M. and Djurovic, P. 2013. Research of Marine
Accidents through the PRISM of Human Factors.
PROMET‐Traffic & Transportation. Vol.25 No.4: 369‐
377.
Endsley, M.
R. 1995. Toward a Theory of Situation
Awareness in Dynamic Systems. Human Factors: The
Journal of the Human Factors & Ergonomics Society.
Vol.37Issue1:32‐64.
Grech,M.,Horberry,T.andSmith,A.2012.HumanErrorin
Maritime Operations: Analyses of Accident Reports
UsingtheLeximancer Tool,Proceedingsof
theHuman
FactorsandErgonomicsSocietyAnnualMeeting.Vol.46
No.19:1718‐1721.
Hara,K.andNagasawa,A. etal.1990.TheSubjectiveRisk
AssessmentofShipsCollision.JournalofJapanInstitute
ofNavigation.Vol.83:71‐80.(inJapanese)
Hara, K. and Nakamura, S. 1995. A Comprehensive
Assessment System for the Maritime
Traffic
Environment.SafetyScience.Vol.19Issue2‐3:203‐215.
Koester, T. and Sorensen, P. K. 2003. Human Factors
Assessment.ProceedingsofInternationalConferenceon
Marine Simulation and Ship Maneuverability 2003. I.
RA‐20.
Nishizaki,C.andItoh,H.2015.DevelopmentofaMethod
for Ship Collision Analysis with Bridge
Simulator.
International Journal on Emerging Trends in
EngineeringandTechnology.Vol.3.Issue1:11‐20.
Nishizaki, C and Takemoto, T. 2016. Measurement of a
Navigatorʹs Situation Awareness for Crossing Ships
using SAGAT. Proceedings of Asia Navigation
Conference2016.7‐13.
Okazaki, T. and Ohya, M. 2012. A Study on Situation
Awareness of Marine Pilot Trainees in Crowded Sea
Route.ProceedingsofIEEEInternationalConferenceon
System,ManandCybernetics2012.1525‐1530.
Romer,H.&Petersen,H. J.S.&Haastrup,P.2009.Marine
Accident Frequencies‐Review and Recent Empirical
Results‐.JournalofNavigation.Vol.48:410‐424.
.
