271
1 INTRODUCTION
The International Maritime Organization (IMO) is
almost reaching to the finalization of the second
generation of intact stability criteria (IMO 2017). It
will be finalized by the Sub‐Committee on Ship
DesignandConstruction(SDC)till2019.Thecriteria
aredefinedinthreelevels:(Level1)thelargestsafety
ma
rgin but simplest calculation method, (Level 2) a
medium safety margin but more complicated
calculation method, and (Level 3) quite complicated
direct stabilityassessment with the minimum safety
margin.Ifashipfailstopassthefirstandsecondlevel
criteria, the ship should satisfy the direct stability
assessmentor change thecargo loading condit
ionto
adapt the criterion. Otherwise, the ship should be
operatedwithoperationallimitationswhichhavenot
been handled in the current intact stability criteria.
Theintroduction ofoperational limitations is agreed
uponatthesub‐committeeofSDCinprinciple.
Under the current intact stability crit
eria, a
shipmaster needs to confirm whether the ship with
current/expected loading condition satisfies the
stability requirements in the regulation or not. If
cleared, the shipmaster can sail the ship anywhere
without any limitations. However, selection of
navigation routes is limited according to the
operational limitations in the second generation of
int
actstabilitycriteriatoguaranteethesafetyofships.
Therefore, impact assessment of the operational
limitations on actual ship operation should be
carefullyinvestigatedintermsofimplementation.
Nowadays, weather routing algorithms are
developedasfordecision‐makingtoolsformastersto
select an economical route with safety. Various
weather routing algorit
hms for ship safety has been
proposed(e.g.Krata&Szłapczyńska,2012).Inorder
todiscusstheoperationallimitations,wetrytoassess
influence of the operational limitations using a
Analysis of Satellite AIS Data to Derive Weather
Judging Criteria for Voyage Route Selection
M.Fujii
M
arineTechnicalCollege,JapanAgencyofMaritimeEducationandTrainingforSeafarers,Ashiya,Japan
H.Hashimoto&Y.Taniguchi
GraduateSchoolofMaritimeSciences,KobeUniversity,Kobe,Japan
ABSTRACT: The operational limitations are discussed at the IMO as a part of the second generation intact
stabilitycriteria.Sinceitisafirstattempttointroduceoperationaleffortsintosafetyregulations,comprehensive
discussions are necessary to realize practically acceptable ones. Therefore thi
s study investigates actual
navigationroutesofcontainershipsandpurecarcarriersinthetrans‐NorthPacificOceaninwinter,because
theyarepronetosuffersignificantparametricrollwhichisoneofstabilityfailuremodes.Firstly,interviewsare
madetoshipmasterswhohaveexperiencestohaveoperatedthesubjectshipstoident
ifymajorelementsfor
routeselectionintheNorthPacificOcean.Secondly,sufficientnumberofactualnavigationrecordsiscollected
fromSatelliteAISdatatoderivetheweathercriteriafortherouteselectioninsevereweathercondition.Finally,
shipmaster’son‐boarddecision‐makingcriteriaarediscussedbyanalysingtheshiptra
ckingdataandweather
data.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 11
Number 2
June 2017
DOI:10.12716/1001.11.02.09
272
navigation simulation for ocean‐going ships
(Hashimoto et al., 2016). The navigation simulation
modelwasdevelopedbasedonaweather‐routingone
developedbyKobayashietal.(2011,2015).However
actual route‐decision process of ships in severe
weatherisnotclearenoughwhereasitisoneofmost
important
elements for the development of rational
operationallimitations.Inthiscontext,thenavigation
simulation is not only a simulation tool for route
suggestiontocaptains,buttochecktheadaptationto
the new stability criteria. Therefore, the navigation
simulationshouldsimulateshiprouteswithsufficient
similarity with actual routesdecided
by captains by
taking account of the preferred safety margin
depending on the weather condition/forecast. The
route decision‐making criteria for regulatory
purposes must be objectively. There are several
researches in this direction (Vettor & Soares, 2015;
Hayashi & Ishida, 2004), but they are not objective
satisfactorily.
Byfollowingthesesituations,
thisstudyconducts
aninvestigationofshipactualroutesofshipsserved
inTrans‐northPacificinwinter.Thesubjecttypesof
shiparecontainer ships andpure carcarriers (PCC)
who are assumed to be affected by the second
generationintactstabilitycriteria.Firstly,shipmasters
whohave operated container
ships and/or PCCsare
interviewed to determine the route selection criteria
of the trans‐North Pacific Ocean routes in winter.
Secondly,AutomaticIdentificationSystem(AIS)data
receivedbysatellitesareusedasoneofmostobjective
data to reveal actual navigation routes in severe
weather condition to describe the relationship
between
the route selection and weather judging
criteria. Finally, the shipmaster’s on‐board decision‐
making criteria are discussed by analyzing the
trackingdataofactualshipsandweatherdata.
2 SHIPMASTERON‐BOARDROUTEDECISION
The developed navigation simulation is based on a
weather routing model. Even during actual
navigation, a
weather routing service is commonly
used; this means that weather routing is a main
method for correctly simulating practical navigation
routes. However, the route decision is ultimately
decided by the shipmaster. Hence, the shipmaster’s
intentionsneedtobeincludedintheweatherrouting
model.
In this study, shipmasters who were experienced
withacontainershipand/orPCCwereinterviewedto
determinetheweather criteria and limitationsof the
trans‐North Pacific Ocean route in winter. Here,
criteriarefertoa standardfornavigationwithoutany
restriction,and a limitation refers toa standard that
does not allow navigation. The findings of
the
interviewsarediscussedbelow.
2.1 Routeselection
Theshipmasterbasicallyselectsa routeaccordingto
the minimum distance and least ship motion
consideringthelocationoflow‐pressureareas.Along
the eastbound route from Asia to North America,
shipmasters navigate by great circle sailing. The
southernpartofthe great
circleisselectedto utilize
the tailwind and following waves from a low‐
pressure.Inaddition,shipmastersgenerallyselectto
navigatebehindalow‐pressureintheGulfofAlaska.
AlongthewestboundroutetoAsia,shipmastershead
towardstheBeringSea,where thewindsandwaves
arecalm. If
itisimpossibletoheadnorth,asouthern
routeisselectedbyMercatorsailing.
2.2 Informationusedforrouteselection
A shipmaster decides a route based on the weather
forecast a week before sailing. At that time,
navigationrecordsofpastvoyagesandthepilotchart
arealsoreferredto.Recently,
recommendationsfrom
weather routing service are used; even in that case,
theforecastaccuracyiscarefullyconsidered.
2.3 Effectsofwindand waves
The wave height criterion for container ships is
normally 5 m. Shipmasters select a route where the
forecast wave height does not exceed 6 m, but
sometimes
navigateareaswherethewaveheightis7–
8m.
In the case of PCCs, shipmasters feel that
navigationisdifficultwhenthewaveheightexceeds4
mbecauseofconcernsoverengineperformanceand
shipmotions,e.g.rollingandpitching.Inaddition,a
PCCisaffectedbywindbecauseof
itslargereceiving
area.
During racing and/or torque‐rich are occurred,
shipmasters consciously reduce the ship. Thus, the
receiving direction of the wave is determined by
considering the influence of the ship’s speed and
motion. Shipmasters normally avoid waves from
deadaheadasmuchaspossible.
3 EXAMINATIONOFCONCRETECRITERIA
USINGSATELLITEAISDATA
Basedontheaboveinterviewresults,thecriteriaand
limitationsforthetrans‐NorthPacificOceanroutein
winterwerelooselydetermined.However,thecriteria
needtobedefined more concretelyand numerically
in order to add an algorithm that represents route
decision‐makingbythe
shipmaster.Therefore,criteria
weredevelopedbyinvestigatingthetrackingdataof
actual ships and weather data in addition to the
interviewresults.
Presently, tracking data can be obtained from
satellite‐based AIS or the Long Range Identification
and Tracking system (LRIT). In this study, the
satellite‐basedAISdatawereused,
becausetheLRIT
dataaremostlyusedbygovernments,andtheformer
ismorefocusedoncommercialuse(Chen,2014).
The satellite AIS data were obtained from
exactEarth.Thedataincluded100containershipsand
84PCCsthatwerepickedatrandomfromthevessels
thatcrossedtheNorthPacific
Oceanfrom1
December
2015 to 29
February 2016. The weather and sea
273
conditionswere analysed by usingNational Centers
forEnvironmentalPrediction(NCEP)data.
3.1 Interpolationofpositiondata
Theintervalsofthereceived AIS data varied from a
fewsecondstoafewhours.Thus,intheanalysisthe
positionwasestimatedevery3hstartingfrom00:00
UTCascalculated
fromtheclosestposition.
The mesh of the weather data provided by the
NCEP had longitudinal intervals of 1.25° and
latitudinal intervalsof 1.0°. Linearinterpolationwas
performed to calculate the weather data at a ship’s
positionatagiventime.
3.2 HandlingerrorsinsatelliteAISdata
As shown
in Figure 1, the AIS position data
sometimesjumpedtoanunreachablepoint.Suchdata
neededtobeexcludedfromtheanalysis.Thereason
forthepositionjumpcouldnotbedetermined.Thus,
inthisstudy,adistancebetweenneighbourpositions
wasusedtojudgeifanerrorhadoccurred.Figure
2
showsthesystemflow.
Figure1.ErrordataincludedinsatelliteAISdata.
Figure2.Systemflowforerrorjudgement.
The distance is defined by the ship’s speed data
(SOG) transmitted by AIS from each vessel. Table 1
presents the number of corresponding data, average
speed, and standard deviation, and Figure 3 shows
theprobabilitydensitydistributionofSOG.AllSOG
data obtained for this study were calculated except
whenthe
speedwasbelow0.5kn,whichmeansthat
the vessel was not sailing. As indicatedin Figure 3,
the SOGs followed an almost normal distribution.
Therefore,thelimitdistancewasdefinedfrom 3σ of
SOG data. This means that the AIS data were
regarded as an error if the speed
between neighbor
positions was greater than the limit speed which is
calculated by the data received time and the limit
distance.
Table1.DetailsofSOGdata.
_______________________________________________
Numberofdata Mean(kn) SD3σ
_______________________________________________
Container 3,505,44517.93 3.19 27.49
PCC1,560,48715.98 2.89 24.66
_______________________________________________
Container ship
PCC
Figure3.Probability densitydistribution ofSOGfromAIS
data.
4 DATAEXTRACTIONANDDEFINITIONOFTHE
TRANS‐NORTHPACIFICOCEANROUTE
The AIS data obtained in this study included ships
that navigated the North Pacific Ocean during the
period in question, even if it was only once. This
means that vessels that navigated the North Pacific
Oceanonlyonceand
thennavigatedotherareaswere
included in the data. Therefore, navigation data for
the target area(north of 10° N, between130° E and
110°W)wereextracted.Inordertoanalyseeastbound
andwestboundvoyages,departingandarrivinglines
weresetontheAsianandAmericansides,asshown
inFigure4.
In general, when a vessel navigates the North
Pacific Ocean from the Asian side to the American
sidebygreatcirclesailing,thedeparture point from
the great circle is set off Japan. Therefore, the limit
lineontheAsianside(i.e.west‐sideline)wasdefined
at
143.5° E. Thelimit line onthe American side(i.e.
east‐sideline)wasdefinedasfrom60°N,140°Wto0°
N,112.5°WoffcontinentalAmericabecauseportsare
located widely distributed from the northwest to
274
southeast.However,vesselsmaynavigatejustonthe
east‐sidelinebecausethelinewasdrawndiagonally.
Hence, the limit line was divided into latitudinal
intervals of 1°, and the longitude was calculated at
everylatitude.Thecalculatedlongitudelineevery1°
wasusedastheeast‐sidelimitline
forjudgement.
In this study, eastbound vessels were defined as
passing the west‐sideline first and then passing the
east‐side line. Westbound vessels did the opposite.
Theanalysisofthevoyagesusedvaliddatabetween
the AIS data which was first received after a vessel
entered the area
and the AIS data that was first
receivedafterthevesselleftthearea.
Figure4.Departingandarrivinglines.
5 ANALYSISRESULTS
5.1 RoutecomparisonforcontainershipsandPCCs
Table 2 presents the number of trans‐North Pacific
Ocean voyages during this period. The numbers of
voyages by container ships and PCCs are different
becausethecontainer ships operatedthesame route
according to a schedule, but the route
of the PCCs
wasnotfixed.
Table2.NumberofvoyagesacrosstheNorthPacificOcean.
_______________________________________________
Eastbound WestboundTotal
_______________________________________________
Container164111275
PCC221537
_______________________________________________
Figure5.Plottedroutesforeachtypeofship.
Figure 5 plots the eastbound and westbound
routesofeachtypeofship.Whentravelingeast, both
containershipsandPCCsnavigatedthecentreofthe
NorthPacificaroundalatitudeof40°N.Ontheother
hand, when traveling west, the ships followed two
routes:throughtheBeringSeaand
around30°N.In
addition, the eastbound routes of the both types of
shipshowedalmostthesametrend.However,forthe
westbound route south of 40° N, the PCCs headed
muchfurthersouththanthecontainerships.
5.2 Encounteredwavesanddirection
Table 3 presents the average wave height.
The
containershipsencountered anaveragewaveheight
0.5 m higher than the PCCs. Figure 6 shows the
probability density distribution of the wave height.
The wave distribution of the container ships slowly
decreasedfrom3mto8m.However,thedistribution
decreased sharply for the PCCs. This is consistent
with the interview results, i.e. navigating in areas
withwaveheightsofover4mbyPCCisdifficult.
Table3.Averagewaveheight.
_______________________________________________
Numberofdata Mean(m) SD2σ 3σ
_______________________________________________
Container
Alldata 21,705 3.53 1.29 6.11 7.40
Eastbound 12,423 3.74 1.26 6.25 7.50
Westbound 9,2823.26 1.28 5.83 7.11
PCC
Alldata 3,5013.05 0.97 4.98 5.95
Eastbound 2,1413.16 0.98 5.12 6.10
Westbound 1,3602.86 0.92 4.70 5.62
_______________________________________________
Container
ship
PCC
Figure6. Probability density distribution of the wave
height.
275
Container
ship
‐Head
Container
shi p
‐Abeam
Container
ship
‐Aft
n
=
1,004
Mean
=
2.83m
(SD:0.84)
n
=
735
Mean
=
2.97m
(SD: 0.96)
n
=
1,762
Mean
=
3.20m
(SD: 1.01)
PCC
‐Head
PCC
‐Abeam
PCC
‐Aft
n
=
3,934
Mean
=
3.22m
(SD: 1.29)
n
=
12,250
Mean
=
3.72m
(SD: 1.26)
n
=
5,521
Mean
=
3.34m
(SD: 1.28)
Figure7.Probabilitydensitydistributionofthewaveheight
foreachencounterdirection.
According to the interview results, the wave
heightcriterionis5mforcontainerships,and4mfor
PCCs. These figures indicate an average speed of
almost1σforeachvesseltype.
Figure7showstheprobabilitydensitydistribution
of the wave height for each encounter direction. A
value
of 0° means dead ahead. ‘Head’ refers to a
range between 60° starboard and 60° port from the
bow.‘Abeam’referstoarangeof60°totheheadand
60° to the aft on both the port and starboard sides.
‘Aft’referstoarangeof60°toeachside
fromdirectly
aft. For the container ships, the trendin the density
wasalmostthesameinalldirections,andthedensity
decreased with an increasing wave height. The
highestaveragewave height camefrom the aft,and
lowest came from the head. For the PCCs, the head
and aft
densities decreased sharply above a wave
height of 3.5 m. However, the aft waves showed a
graduallydecreasingtrend.Thewaveheightsforthe
PCCswerehighestfromtheaftdirectionandlowest
fromtheheaddirection.
For both container ships and PCCs, the average
wave height from the head was
5% less than the
averageheightforallvoyages,andtheaveragewave
heightfromtheaftwas5%greater.
Table 4 presents the proportion of the encounter
direction for each wave height category. The
proportion of wave heights from the head suddenly
decreasedover7m,andtheabeam
wavesincreased
instead.Inaddition,thetotalnumberofdataover7m
decreased.Thus,wavescomingfromtheheadshould
belimitedtoaround7m.Inthe case of PCCs,head
waves should belimitedto around5 m becausethe
totalnumberofdatawaslessover5
mthanbelow5
m.
5.3 Eastboundand westboundtrends
Table5presentstheaveragespeeddataforeastbound
and westbound ships, and Figure 8 shows the
encounterwavedirection.AsshowninFigure8,both
container ships and PCCs mainly encountered aft
waves when eastbound. On the other hand,
westboundshipsencounteredwavesfromdiagonally
in front. The container ships received waves from
both port and starboard; however, the PCCs only
received waves on the starboard side. This may be
because of the difference in voyage areas. We
analysed trends for four areas in the North Pacific
Ocean,asshown
inFigure9.Area2receivedwaves
diagonally from the port front, and other areas
received waves from the starboard front side. That
means that westbound container ships may not
navigate in a zigzag fashion to avoid bow waves
basedononlythesedata.
Table5.Averagespeedforeastboundandwestboundships
_______________________________________________
Mean(kn) SD
_______________________________________________
Container Eastbound 18.742.60
Westbound16.263.21
PCCEastbound 15.792.49
Westbound16.692.42
_______________________________________________
Table4.Proportionofencounterdirectionbywaveheight.
___________________________________________________________________________________________
Waveheight(m)≤22–3 3–4 4–5 5–6 6–7 7–8 8–9
___________________________________________________________________________________________
Container
Head %33.51 29.02 23.42 20.83 23.52 22.40 12.98 NA
#707 1,777 1,420 903 495 157 170
Abeam %30.90 19.37 18.11 12.55 14.16 12.41 18.32 40.00
#652 1,186 1,098 544 298 87244
Aft %35.59 51.61 58.46 66.62 62.33 65.19 68.70 60.00
#751 3,160 3,544 2,888 1,312 457 906
PCC
Head %35.59 33.64 25.80 18.66 5.10 22.22 16.67 NA
#147 479 289 755610
Abeam %20.82 22.12 21.79 14.18 14.29 29.63 66.67 NA
#86315 244 5714840
Aft %43.58 44.24 52.41 67.16 80.61 48.15 16.67 NA
#180 630 587 270 791310
___________________________________________________________________________________________
276
Container ship ‐Eastbound
PCC‐Eastbound
Container ship ‐Westbound
PCC‐Westbound
Figure8. Probability density distribution of the wave
directionforeastboundandwestboundships.
Figure9.TheNorthPacificOceandividedintofourareas.
Figure10plotstheshippositionforwaveheights
over5mforcontainershipsandover4mforPCCs.
These wave heights were selected based on the
opinions of shipmasters.The eastbound voyages are
widely distributed. On the other hands, the
westbound voyages mainly distributed around the
Gulf of
Alaska. Here it can be seen that Eastbound
voyagesthatmainlyreceiveaftwavesareallowedto
navigateforalongstretchoftimeevenunderrough
seaconditions.However,thewestboundvoyagesthat
receiveheadwavescouldbeallowedtonavigateonly
ashorttime.Itcanbedetermined
thatthenavigation
periodforhighwavesisrelatedtotheshipmotion.
Figure10. Plotted ship positions for >6 m wave height
encountered by container ships and >4 m wave height
encounteredbyPCCs.
PCC
Container
ship
n
=
3,326
Mean
=
4.09m
(SD: 1.26)
n
=
257
Mean
=
3.51m
(SD: 1.22)
Figure11. Wave height in area north of 30°N and east of
165°W.
Figure11showsthedensityofencounteredwave
heightsinthearea northof30°Nandeastof165°W.
Theaveragewaveheight was 0.5m higherthanthe
average height for all voyages. For PCCs, the high
densityzonereachedaround 5m.Forcontainerships,
the density
was not significantly reduced. However,
thewaveheightincreasedasawhole.Ifthecriterion
is1σoftheaveragespeedlikeotherareas,itwouldbe
around 5.5 m.Thus, the shipmasters’ wave criterion
may have become more lenient temporarily during
navigation of this area. The criterion can be
determinedtobe5.5mforcontainershipsand5mfor
PCCs.
5.4 Relationbetweenthewaveheightandshipspeed
During the actual operation of merchant ships, the
speed may be decreased depending on the sea
conditions. This is decided by the shipmaster
dependingonthemainengine
performanceandship
motion. According to the interview results, a
shipmaster decreases the engine revolution when
racingand/ortorque‐richconditionsareexpected.
Figure 12 shows a scatter plot of the relation
between the wave height and ship speed. For a
container ship that received an aft wave, the ship
speed
stayedalmostthesame(r=0.0202467)evenat
highwaveheights.Ontheotherhand,foracontainer
ship that received a head wave, the ship speed
gradually decreased (r =−0.2780084) as the wave
heightincreased.
277
PCC ‒ Aft waves
PCC - Head waves
Container ship - Aft waves
Container ship - Head waves
r = 0.0202467
r = -0.2780084
r = -0.1187477 r = -0.3745359
Figure12. Scatter plot of the relation between the wave
heightandspeed.
In the case of PCCs, the ship speed gradually
decreased with aft waves (r =−0.1187477) even at
large wave heights. The speed with head waves
decreased (r =−0.3745359) with an increasing wave
height.Theseresultsincludetheeffectofthecurrent,
dirt on the bottom, winds and waves,
and reduced
enginerevolution. Hence, itisdifficult to clarifythe
reason why a ship’s speed decreases. However, the
deceleration rate of PCCs was much greater than
container ship. This result agrees with the
shipmasters’opinionthatthespeedofaPCCshould
beimmediatelydecreasedinroughseas.
6
CONCLUSION
Inthisstudy,trackingdataofactualshipstraversing
the trans‐North Pacific Ocean route in winter were
analysed to determine the relationship between the
route selection and weather judging criteria, with a
particular focus on waves. Based on the results, the
followingweatherjudgingcriteriaweredetermined:
Based
onthe trackingdata and interviewresults,
theaveragewaveheightencounteredbycontainer
ships is 3.53 m; the wave criterion for the North
Pacific Ocean in winter should be 5 m, which is
almostthesameas1σ.
TheaveragewaveheightencounteredbyPCCsis
3.05
m. The wave criterion for the North Pacific
Oceaninwintershouldbe4m,whichisalmostthe
sameas1σ.
For both container ships and PCCs, the average
heightsofheadwere5%lessandaftwaveswere
5% greater, compared with the average wave
heightfor
allvoyages.
Headwavesforcontainershipsshouldbelimited
to around 7 m. Head waves for PCCs should be
limitedtoaround5m.
Eastboundvoyagesthatmainlyreceiveaftwaves
areallowedtonavigateforalongstretchoftime
even under rough sea conditions. However,
westboundvoyagesthatmainlyreceiveheadwave
couldbeallowedtonavigateonlyashorttime.
Intheareanorthof30°Nandeastof165°W,the
average wave height is 0.5 m higher than the
average height for all voyages. In addition, the
high density zone reaches
around 5.5 m for
container ships and 5 m for PCCs. The
shipmasters’ wave criterion may become lenient
temporarilywhennavigatingthisarea.
PCCs decelerate much more under head sea
conditionsthancontainerships.
Theresultsofthisstudyarebasedon3monthsof
AISandweatherdata.
Thus,thecriteriamaydifferat
other times of the year depending on the changing
climate.However,atpresentthereisnostandardfor
developing an algorithmto representa shipmaster’s
on‐board decision‐making. These results can
contribute to the development of the operational
limitationsinthesecondgenerationof
intactstability
criteria. It remains as a future work to propose an
algorithmthatrepresentstheroutedecision/selection
by shipmasters with the preferred safety margin for
the navigation simulation based on the outcomes of
this study. To assess the impact of operational
limitationson actualshipoperation isexpected as
a
futureworkusinganupdatednavigationsimulation
basedontheoutcomesofthisstudy.
ACKNOWLEDGMENT
Thiswork was supported by the research activityof
the Goal‐Based Stability Criterion Project of Japan
Ship Technology Research Association in the fiscal
yearof2016,fundedbytheNipponFoundation.This
studywas
alsopartlysupportedbytheFundamental
Research Developing Association for Shipbuilding
and Offshore. We would like to thank Editage
(www.editage.jp)forEnglishlanguageediting.
REFERENCES
IMO.2017.SDC4/WP.1.
Krata, P., & Szłapczyńska, J. 2012. Weather hazard
avoidance in modeling safety of motor‐driven ship for
multicriteriaweatherrouting.TransNav,theInternational
Journal on Marine Navigation and Safety of Sea
Transportation6(1):71–78.
Hashimoto,H.,Taniguchi,Y.,&Fujii,M.2016.Astudyon
operational limitation of the second generation intact
stability criteria. In Proc. of the Japan Society of Naval
ArchitectsandOceanEngineers,Okayama,21–22November
2016.Tokyo:JASNAOE.(inJapanese)
Kobayashi,E.,Asajima,T.,&Sueyoshi,N.2011.Advanced
navigationrouteoptimizationforanoceangoingvessel.
TransNav, the International Journal on
Marine Navigation
andSafetyofSeaTransportation5(3):377–383.
Kobayashi,E., Hashimoto,H., Taniguchi,Y., &Yoneda,S.
2015. Advanced optimized weather routing for an
ocean‐goingvessel.InProc.ofthe2015Int.Associationof
Institutes of Navigation World Congress, Prague, 20–23
October2015.NewYork:IEEE.
Vettor,R.,
&GuedesSoares,C.2015.Detectionandanalysis
of the main routes of voluntary observingships in the
NorthAtlantic.TheJournalofNavigation68:397–410.
Hayashi,M.,&Ishida,H.2004.Weatherroutingsimulation
of oceangoing ship by practical navigators and
encountered wind and wave conditions on simulated
ship’s routes.
The Journal of Japan Institute of Navigation
110:27–35.(inJapanese)
Chen,Y.2014.Satellite‐basedAISanditscomparisonwith
LRIT. TransNav, the International Journal on Marine
NavigationandSafetyofSeaTransportation8(2):183–187.
