397
1 INTRODUCTION
Themaritimetransportationsectorisunderpressure
to make improvements in several areas. Firstly, the
environmental footprint must be reduced to achieve
thegoalof70%reductionofgreenhousegasemissions
by 2050 [1].Secondly, the sector is facing increasing
challengesintherecruitmentofseafarers.Thirdly,the
sector needs to increase safety, both in terms of the
crewandvaluableassets.Fourthly,thesectorneedsto
increasecompetitivenesstomeetthegoalofamodal
shifttowardswaterbornetransportation.Finally,there
is a need to increase resilience in the maritime
transport towards situations such as the COVID‐
19
pandemicwhichdeterioratedseafarersworklifeand
created difficulties with accessing ports. In this
context, the maritime industry is investigating
autonomous ships as an alternative to conventional
ships. Automation can improve safety by removing
people from potentially dangerous tasks like cargo
handling. Another benefit of automation is the
potential to
reduce manning, which can remedy the
shortage of seafarers and reduce the crew cost.
Moreover, for a fully unmanned ship, the
superstructurecan be removed,increasing the cargo
carrying capacity, lowering energy usage, and
enabling new ship designs. By automating different
tasksonaship,theneededmanningcanbe
reduced.
However, operating alarge cargo ship unmanned is
notstraightforward.Oneofthebiggestchallengesfor
operating alarge cargo ship unmanned todayis the
frequent need for maintenance on the power and
propulsionsystemandinternationalregulations.
TheautonomouscontainershipYARABirkeland,
ASKOʹs autonomous vessels, and
MEGURI2040 are
just some of the commercial initiatives that aim to
operate autonomous ships in the near future. There
exists a lot of research on the challenges and
feasibilities of autonomous ships, e.g., [2]–[9].
However, there is less discussion on the transition
from manned to unmanned ships, as pointed out
in
[7]. This study will investigate the technical and
regulatoryfeasibilityofautomatingdifferenttasksfor
Analyzing the Feasibility of an Unmanned Cargo Ship
for Different Operational Phases
P.RøstumBellingmo
1
,E.Wille
2
,H.Nordahl
1
,O.E.Mørkrid
1
&E.A.Holte
1
1
SINTEFOcean,Trondheim,Norway
2
SINTEFÅlesund,Ålesund,Norway
ABSTRACT:Themaritimeindustryhasbeguntolookintoautonomousshipsasanalternativetoconventional
ships due to growing pressure to reduce the environmental impact of maritime transportation, to increase
safety,tomitigatethegrowingchallengesinrecruitingseafarers,andtoincreaseprofitmargins.There
isalotof
researchonthe challengesandfeasibilitiesof anautonomous ship.However,thereis lessdiscussionon the
transitionfrommannedtounmannedshipsandthetasksthatarefeasibletoautomatebeforethewholeshipis
unmanned. This paper investigates the technical and regulatory feasibilityof automating
different tasks for
differentoperationalphasesforalargecargoship.Thisstudyshowsthatafullyunmannedcargoshipisnot
feasibletoday,butthatsometaskscanbeautomatedwithinthenextfiveyears.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 17
Number 2
June 2023
DOI:10.12716/1001.17.02.1
7
398
differentoperationalphases.Assuch,itwillserveasa
guidancetotakingthestepstowardsunmannedships
by identifying what autonomous operations are
feasibletoday,andwhichremainingchallengesmust
besolvedtotakethefinalsteps.Thescopeislimited
to unmanned cargo ships operating internationally,
e.g.,ships
operatingwithinEuropeantradeorinthe
deep sea segment. This paper uses the term
unmannedshiptorefertoashipwithoutanypeople
onboardandoperatingunderconstrainedautonomy.
A constrained autonomous ship is defined as an
ʺuncrewed operation with constrained autonomy
onboard but with operators in
Remote Operation
Center (ROC) that can handle more complex
situations.ʺ[10].
2 METHOD
The methodology in this study follows the work in
[11]. The study investigates the technical and
regulatoryfeasibilityofanunmannedcargoship.The
evaluationisperformedbylookingintospecifictasks
and operational phases. The tasks
and phases have
been determined based on feedback from shipping
companies.Theoperationalphasesinclude1)atport,
theshipisstationaryataport,2)nearport,theshipis
arrivingorleavingaport,3)coastal,theshipis ina
highlytraffickedareaornarrowwaters,and4)
deep
sea, the ship is in open sea. This study covers the
following tasks: navigation, propulsion,
communication, cargo handling, and mooring. To
narrowthe scopeof thisstudy, additionaltasks that
arerequiredtooperateacargoshiparenotincluded.
Atrafficlightisusedtoindicatethefeasibility
with
the colors defined in Figure 1. A color of feasibility
status is given for each task for each operational
phase.
Figure1:Colorsdefiningfeasibility.
If all the tasks can be performed at all the
operational phases, the ship can operate unmanned.
The assessment of feasibility is based on a
combination of literature study and semi‐structured
interviews with technical and service providers,
authorities,and classificationsocietyinthemaritime
industry.Detailsonthiscanbe
foundin[9],[11].
3 TECHNICALFEASIBILITY
In terms of evaluating the technical feasibility of
realizingunmannedcargoshipoperations,thispaper
takes a closer look at navigation, propulsion,
communication,mooring,andcargohandling.
3.1 Navigation
Navigation is the act of getting a ship from one
locationtoanotherin
asafeandefficientmanner.To
navigateashipwithnocrewonboard,itisnecessary
toestablishsituationalawareness,i.e.,observe,detect,
and classify objects in the surroundings of the ship
and predict the behavior of the objects. Next, the
navigationsystemmustplanapathfortheunmanned
shipandsteertheshipaccordinglywhileensuringthe
safety ofthe ship, cargo, and the environmentof its
operation.
3.1.1 AutomaticNavigationinCongestedandConfined
Waters
Today, several solutions for autonomous
navigationareavailableonthemarket,e.g.,[12]–[17],
where some have been tested for large cargo ships
[15], [18], [19]. Based on interviews with leading
technology and service providers in the maritime
industry, one of the biggest hurdles to overcome
towards realizing unmanned navigation of ships is
object classification. It is crucial for the navigation
system to know what type of objects are in the
surroundings,forinstance
ifitisashiporshore.Alot
ofhigh‐qualitydataisneededtodevelopanaccurate
objectdetectionalgorithm,whichiscurrentlylacking
[9].Anotherchallengeisfortheautomaticnavigation
systemto know itsownlimitations,thatis,when to
callforassistancefromanROC
orgointoasafestate.
Thisisalsoachallengeforallotherautomatedtasks.
Moreover, behavioral prediction and path planning
will be a challenge when the unmanned ship must
interactwithotherships[20].
3.1.2 NoPhysicalPilotageonBoard
In specific areas, such as in port areas or
highly
traffickedareas, apilot is required to safely guide a
ship[21].Today,thisoperationrequiresthatthepilot
physically boards the ship. As outlined in [22], this
becomesanissueinthecaseofanunmannedship,as
the pilot will not have a bridge to perform its
tasks
from.Asapossiblesolutiontothisissue,Porathe[22]
suggests a digital pilot service with the same local
knowledge and expertise as a conventional pilot.
However, solutions for digital pilotage are not yet
realized. Another possible solution is for the ROC
operatorstohaveapilotagecertificateforthe
areaof
operation. However, for international voyages, this
canbeachallengeduetothelargeareaofoperation.
3.1.3 AutomaticDockingWithoutAssistanceFrom
Tugboats
When aship approaches a quay, it must prepare
for and carry out the docking operation, i.e., sailing
fromthefairwayareatothe
quay.Whenleavingthe
quay, the operation is reversed, and un‐docking is
performed before the ship returns to the fairway.
Someactorshavealreadydemonstratedautonomous
docking[23]–[25].However,informationontheactual
performance of these docking systems, for example
theoperationalstability,isnotpubliclyavailable.
399
3.1.4 Feasibility
Basedontheabove,itisreasonabletoassumethat
the ability of an oceangoing cargo ship to perform
autonomousnavigationinasafeandefficientmanner
is not feasible today, but within the next 5 years
(givingayellowtrafficlight).Themainreasonisthe
existing
operational limitations of todayʹs
autonomous navigation systems related to object
detectionand classification. Nevertheless,automated
navigationisconsideredlesschallenginginthedeep
sea operational phase compared to the other
operational phases, mainly due to thelimited traffic
andlackofcongestedareas.
3.2 Propulsion
For a conventional ship,
there is crew on board to
handleanypotentialfailuresandmaintenanceneeded
with the propulsion and support system (Nedcon
Maritime, 2013). However, for an unmanned ship,
there is no crew on board, meaning that the
propulsion system must be able to operate without
needforinterventionorrepairsforthe
entirevoyage
(i.e., sailing between two ports). Cargo ships
operatinginternationallyaretypicallyequippedwith
low‐cost fossil fuelpropulsion systems, e.g.,a diesel
engine using VLSFO, which requires a lot of
maintenance. However, there exists alternative
propulsion systems and cleaner fuels which require
lessmaintenance.Forshortervoyages(e.g.,
15‐30nm
asforYaraBirkeland),electricalengineswithbatteries
can be used, which require less maintenance.
Unfortunately, for cargo ships operating long
distances,maintenancefreepropulsionsystemshave
alowleveloftechnicalmaturity.Thereexistsolutions
for maintenance and fault prediction, however, this
would still require a power
and propulsion system
thatcouldoperatewithoutfailuresbetweenports.
3.2.1 Feasibility
Automated propulsion systems for cargo ships
sailing long voyages are not considered feasible
withinthenext5years,duetothehumanintervention
needed with todayʹs propulsion system. Further
development of new propulsion systems without
need for
maintenance on long voyages is necessary.
However, for shorter voyages, where for instance
electricalenginescanbeused,automatedpropulsion
systemispossiblewithinthenext5years.Thus,ared
traffic light is given for the feasibility in deep sea,
while a yellow traffic light is given for the at
port,
near port and coastal phases due to the shorter
distances.
3.3 Communication
For a cargo ship, communication with other ships,
ports,andauthoritiesisrequired.Additionally,foran
unmanned ship, communication with an ROC is
needed.Movingtheresponsibilityfromtheshiptoan
ROCcreatesnewrequirementsfor
communication.
3.3.1 GainSituationalAwarenessAroundtheUnmanned
Ship
Communication requirements for an ROC
operatinganunmannedshiparenoteasilyavailable
as this is still under development. Operators at an
ROCneedtogainsufficientsituationalawarenessto
safely operatethe unmannedship. A challengeis to
determine the needed
data to gain situational
awareness, which is limited by the available
bandwidth. Thus, the ROC must prioritize what
should be communicated. Regardless, the ROC will
utilize all available bandwidth. The communication
bandwidth available depends on where the ship is
located. Near shore, terrestrial communication with
high bandwidth is typically available.
In deep sea,
only satellite communication is available, which has
limitedbandwidth.Lowbandwidthlimitstheamount
ofsensordatathatcanbetransferredtoanROCfrom
an unmanned ship. This means that it might not be
sufficient bandwidth to gain full situational
awarenessattheROCinsomeareas.
Inthenextyears,
satellitecommunicationsystemswithhighercapacity
and lower latency are expected, such as Starlink.
Moreover,asthelevelofautonomyrisesontheship,
theamountofdataneededinanROCisexpectedto
decrease.
Anotherchallengeistoensurethedataqualityand
reliability,
as faulty data can impact the safety and
efficiencyoftheoperation.Ifcommunicationwiththe
ROC is lost, the ship should enter a minimum risk
condition if the situation is challenging. Further
information on communication challenges and
possibledatarequirementsforanROCisdiscussedin
[27],howeverthestudy
issomeyearsold.
3.3.2 PossibletoRemotelyControltheUnmannedShip
The ROC must be able to remotely control the
unmanned ship if the automation system cannot
handle the situation. Low latency is crucial when
operatingtheshipremotely.Whentheshipiscloseto
shore, there are communication
systems with low
latencyavailable,e.g.,5G.Whenfarfromshore(i.e.,
deep sea), only satellite communication is available,
whichhas ahighlatency.However, when operating
deepsea,decisionswillbelesstimecriticalandthusa
largerlatencycanbeaccepted.Asmentionedearlier,
when Starlink is up and
running globally,
communicationwithlowlatencyshouldbeavailable.
3.3.3 CommunicationWithOtherShips,Authorities,and
PortsinReal‐Time
Evenforanunmannedship,communicationwith
nearby ships, ports, and authorities is needed. All
shipsmustfollowtheConventionontheInternational
Regulations for Preventing Collisions at Sea
(COLREGs),
i.e.,therulesoftrafficatsea.However,in
sometrafficsituationstheserules arenot expliciton
whatactionshouldbetaken,whichmakesithardfor
the automated navigation system to determine the
proper action. In such situations, ships must
communicate with each other anddecide who takes
what
action (to avoid a collision). Until new
COLREGs are in place, communication with other
ships must be handled by the ROC [20]. Route
exchangebetweenshipshasbeenproposedasanother
400
possible solution to avoid unclear traffic situations.
However,forthistowork,alloperatingshipsmustbe
equipped with route exchange systems, which is a
comprehensive task [20]. Even more, with route
exchange systems in place on all ships, there is still
need for explicit rules on how to behave
if these
routesintersect.ThismeansthattheCOLREGsneeds
tobealteredoramendedtomakethemexplicit.
Today, this type of communication is done over
radio.Voicecommunicationfromtheunmannedship
can be transferred to an ROC using Voice over
Internet Protocol (VoIP). It is assumed that
limited
bandwidth is needed to communicate with ports,
authorities,andships.However,thelatencycannotbe
toohigh.This shouldnotbe aproblem accordingto
[28], as for instance Iridium Certus offers voice
communicationwithoutnoticeabledelay.
3.3.4 Feasibility
For the phases at port, near port, and coastal,
terrestrial communication
systems are typically
available with high bandwidth and low latency,
allowingthe ROCto gain situational awareness and
remotelycontroltheship,givinga greentrafficlight.
In deep sea, gaining full situational awareness and
remotelycontroltheshipcanbeachallengeduetothe
limitedbandwidthandhigh
latency.However,foran
unmanned cargo ship, communication with an ROC
cannot be safety critical as the ship must maintain
safetyintheeventofacommunicationsystemfailure
(itmusthavefallbacksolutions).Moreover,duetothe
expectedimprovementsofsatellitecommunicationin
thenextyears,e.g.Starlink,ayellow
trafficlighthas
beengiven.
3.4 Mooring
Traditionalmooringrequirescrewtosecuretheship
usingropes.Withlargeforcesinplace,thistraditional
method poses a large danger for the crew involved.
Whendiscussingmooringinthispaper,itisassumed
that the ship is stationary at port. This
section
summarizes the outcomes from a gap analysis on
automated mooring systems [29] with highlights on
themainchallenges.
3.4.1 AutomaticallyMooringaCargoShip
There exist different solutions to automatically
moor a ship in themarket today, includingvacuum
mooring(e.g.[30]–[32]),magneticmooring(e.g.[30]),
and robot arms (e.g.,
[30], [33]). The vacuum based
mooringsystemisthemostestablishedsolutioninthe
market andin use onmany differenttypes ofships,
rangingfromlargecargoships(<300m)tosmallercar
andpassengerferries.Thetypicalinstallationconsists
ofmultiplevacuumpadsinstalledatport,wherethe
number of pads depends on the size of the ship it
should be able to moor. The automated vacuum
mooringsystemcanreducethemooringtimeandthe
human risk compared with manual mooring. A
downside of the vacuum mooring is that it requires
installationsateveryporttheshipvisits.
Therobotic
armmooringsystem, ifplaced ontheship,canvisit
any port without requiring any other on‐site
infrastructure than the conventional bollard.
However, the robotic arm mooring systems are still
under development and have not been extensively
testedduringoperation. Therefore,itis unclear how
wellthis
mooringsystemactuallyperforms.
3.4.2 AutomaticMooringwithDraughtandTidal
ChangesandRollingMotions
Mooring systems must be able to handle tidal
changesanddraughtdifferencewhenloadingcargo.
Tidalchangescanbe upto 7.5min someport sand
shipdraughtcandifferwith12mbetweenan
empty
ship and a fully loaded ship [34]. By mounting the
mooring units on vertical rails, the vacuum and
magnetic automated mooring systems can adjust to
changesintideandshipdraughts.Anotheroptionfor
adapting to varying heights is to use the different
vacuum pads to step along the
hull. The vacuum
based automated mooring systems can handle roll
motionsinducedbycargohandlingupto6°,whichis
adequateformostsituations.
3.4.3 Feasibility
Automated mooring systems are in use on cargo
shipstodayandarethusconsideredfeasible(givinga
green traffic light). There are, however, some
challenges
with the operational stability of todayʹs
systems.
3.5 CargoHandling
Whentheshipisatport,thecargomustbeloadedon
andofftheship.Additionally,itisimportanttosecure
the cargo on board the ship. Cargo ships can carry
many different types of cargo, e.g., containers,
dry
bulk,wet bulk,andbreak bulk,andthedifficultyof
automating the cargo handling varies between the
different cargo types. Today, there exist cranes that
canautomaticallyon‐andoff‐loadcontainers.Thisis
availableonlarge,advancedportstoday,but noton
smallerports.Forcontainers,thebiggestchallenge
is
securingthecontainersonboardtheship[11].When
the containers are stacked above deck, lashing is
required to avoid movement of the containers. The
lashingis manual workand difficultto automate.A
possible solution is to use cell guides to secure the
containers. However, this is not
possible for large
container ships due to the height of the container
stacks. For smaller container ships, like Yara
Birkeland,thisisnotanissuesincethecontainerscan
be stored below deck. For break‐bulk, automated
cargo handling is challenging, particularly the
connection(hooking)ofthecargo[11].Thisis
mostly
due to the wide variety of goods and packaging,
complicating the automation task. There exist some
semi‐automated solutions for connection in the
market,butalackoffullyautomatedsolutionsforany
typeofgoods.Thisisatopicforfurtherresearch.Wet
anddrybulkareeasier
toautomateasthecollection,
transferring, releasing, and securing of the cargo is
simpler.Therealreadyexistsomesolutionsforthisin
themarket,suchas[35].
401
3.5.1 Feasibility
The feasibility to automate cargo handling
dependsonthetypeofcargo.Break‐bulkisdifficultto
automate,whileforcontainersandwetanddrybulk
there already exist some solution for automation.
Thus,ayellowtrafficlightisgivenforthefeasibility
ofautomatedcargohandling.However,
tosailacargo
ship unmanned, it is not strictly necessary to
automate thecargo handling, ascrew can board the
ship at port. However, this does not solve the
challengeofsecuringcargoduringthevoyage.
4 REGULATORYFEASIBILITY
Theregulatoryfeasibilityofanunmannedcargoship
will be
evaluated based on the different tasks
navigation,propulsion,communication,mooring,and
cargohandling.Existingregulationsforapprovaland
operation of ships have been developed under the
basicassumptionthatimportanttasksareperformed
by humans on boardtheship. Although automation
hastakenovermoreandmorehumantasksinrecent
decades,conventionalshipsalwayshaveahumanon
board with the operational responsibility. Hence, a
shifttowardsunmannedshipscreatesgapsinexisting
rules,regulations,andstandards,sincethesetypically
requireexplicithumanpresence.
Current autonomous ship initiatives mainly
involve relatively short routes, such that only
domesticregulationsandauthorities
arerelevant.This
allows the involved actors (developers, authorities,
etc.)tofocusonalimitedandwell‐definedconceptof
operation(CONOPS).Insupportofaʺstandardizedʺ
approval process,some maritime authorities suchas
the NMA have established a case‐by‐case approval
procedure[36].These case‐by‐caseapprovalsare
very
demanding when it comes to documenting how the
autonomous systems perform and how
responsibilities are distributed between the
automationsystemsandROC indifferentsituations.
It is necessary to verify that equivalent safety is
provided with the automation systems and the
process becomes more demanding as the area of
operationincreases.Whenexpandingtheoperationto
arbitrary and global routes, the ship and its
supporting infrastructure (such as ROC), must also
complywithinternationalregulations.Therearealso
certain legal questions which require clarification,
especially when it comes to the jurisdiction of flag
states,ports,andcoastalstates.Currently,there
areno
common international regulations for unmanned
ships,butIMOisworkingonaʹMASSCodeʹwhichis
expected tobe released in the mid 2020ʹs [37]. Until
the MASS Code is released and in force, any
international unmanned voyage will require specific
agreementsbetweentheshipownerand
allinvolved
authorities.
Other conventions, such as the International
Convention on Standards of Training, Certification
and Watchkeeping for Seafarers (STCW) and The
International Search and Rescue Convention (SAR),
havebeenleftouttolimitthescopeofthestudy.
4.1 Navigation
Autonomous navigation is particularly challenging
from a regulatory perspective because
it involves
extensivedecision‐makingandinteractionwithother
seafarers.Rulesfornavigation atseaaredetermined
byCOLREGs,whichhasbeenwrittenforaworldof
manned ships and are not straightforward for
algorithms to interpret. For example, terms such as
ʺgood seamanshipʺ andʺsafe speedʺ require
subjective interpretations
based on context. Another
examplewhichisdifficultforautonomoussystemsto
handleisthattheCOLREGsspecifythatashipmust
givewaytoanothershipifitiscrossingitspath.Itis
notalwaysclearhowashipshoulddetermineifitis
crossing the path of
another ship, especially in
crowdedwaterways.Anotherchallengeistomakean
autonomoussystemʺbehaveʺinawaythatmakesits
intentionscleartootherseafarers.
Whileit isfeasibleto demonstrate anddocument
equivalent safety levels for autonomous navigation
systems for a limited operational domain, it is not
trivialto
getsuchsystemsapprovedforʺfreesailingʺ
between arbitrary ports. Hence, international
operations for unmanned ships are not considered
feasiblewithinthenext5years,indicatingaredtraffic
lightfordeepsea.However,forcoastalandnearport
phases a yellow traffic light is given, as unmanned
navigationalong
specificroutesisconsideredfeasible
onacase‐by‐casebasis.
4.2 Propulsion
Requirementsregardingreliabilityandfaulttolerance
for systems related to propulsion and steering are
particularly demanding to comply with for
unmannedships sailinglongdistances. For instance,
DNVʹsclassguidelineDNV‐CG‐0264forautonomous
ships[38]
mentions,withreferencetoSOLASCh.II‐1,
thatʺMeans shall be provided whereby normal
operation of the propulsion machinery can be
sustainedorrestoredeventhoughoneoftheessential
auxiliariesbecomesinoperativeʺandthatanavigable
speedmustbemaintainedincaseofpotentialfailures
of single systems
or components. DNV‐CG‐0264
further states that unmanned ships should be
arranged with a minimum of two independent
propulsion lines. Based on the strict redundancy
requirements for propulsion systems, a red traffic
lightisgivenforthedeepseaphase,asmeetingsuch
requirementsdoesnotseemfeasibleforlong
journeys
withinthenext5years.Forthecoastalandnearport
phases,however,it shouldalreadybefeasibletoget
case‐by‐case approval for specific routes, while
approval for free sailing is assumed to be realistic
withinthe next 5 years (giving ayellow trafficlight
forthese
phases).
4.3 Communication
According to the regulatory scoping exercise
conducted by IMO [39], the requirements in SOLAS
chapter IV (Radiocommunications) do not take into
account remote operations and on‐shore control
centers,andthemostappropriatewayofaddressing
thisissueistodevelopanewinstrumenttoproperly
402
addressMASSoperations.SuchaMASSinstrumentis
not expected to be in place and in force within the
next5years,givingaredtrafficlightforthedeepsea
phase.Forthenearportandcoastalphases,ayellow
light is given, since limited operation is considered
feasible
basedoncase‐by‐caseapprovalfollowingthe
NMAcircular[36].Fortheatportphase,agreenlight
is given since this phase does not involve any
navigation or maneuvering and any applicable
requirementsare expected tobe met by state‐of‐the‐
arttechnology.
4.4 Mooring
Unmanned mooring
operations do not appear to
introduceanyrequirementswhichcanʹtbe metwith
todayʹs state‐of‐the‐art technology, and mooring is
therefore given a green traffic light for regulatory
feasibility.
4.5 Cargohandling
The regulatory feasibility of unmanned cargo
handling has not been evaluated in detail, partly
because
the main challenges appear to be technical
ratherthanregulatory. However,thefollowingrules
and standards from DNV have been identified as
relevant:
DNV‐RU‐SHIP Pt.5 Ch.1 Bulk carriers and dry
cargoships
DNV‐RU‐SHIPPt.6Ch.4Cargooperations
DNVGL‐ST‐0377 Standard for shipboard
lifting
appliances
The above doesnʹt seem to have specific
requirements for autonomous cargo handling, but a
yellow traffic lightis given sincesuch systems have
notyetbeentested(orapproved),and itisexpected
that regulatory hurdles may be met during
development and approval of unmanned cargo
handlingsystems.
5 RESULTSANDDISCUSSION
Figure 2 gives a summary of the technical and
regulatory feasibility for operating an unmanned
cargoship internationally. The findings ofthis work
indicate that, over the next five years, it will not be
possible to operate an unmanned cargo ship on
internationalwaters.Thema in
challengesarethehigh
frequent maintenance required with todayʹs
propulsion and sub systems and regulatory
challengesforinternationaloperationsofcargoships.
However,specifictasksforspecificoperationalphases
canbeautomatedtodayorwithinthenextfiveyears.
Thiscanreducetheneededmanningforthatphaseor
related
tothespecifictask.Forinstance,automatically
mooring cargo ships is possible today. Moreover,
automatichandlingofsometypesofcargoisexpected
within the next five years. Autonomous navigation
should be technically feasible in deep sea and areas
withlowoperationalcomplexitywithinthenextfive
yearssince the
trafficchallengesare limited. Froma
regulatory point of view, periodically unmanned
navigation should be possible in the near future as
therearelargecargoshipstodaythatareallowedto
operatewithoutproperviewduetolargecargostacks
blocking the bridge. However, exceptions on the
lookout rules need to
be approved for each specific
voyage. Nevertheless, it is not straight forward to
reducetheoverallmanningofacargoship.Thecrew
are typicallyinvolved in multiple different tasks for
different operational phases. An analysis of crew
reductionisdiscussedin[7].
Figure2. Technical and regulatory feasibility of an
unmannedcargoship.
6 CONCLUSION
Beforecargoshipscanoperatecompletelyunmanned
internationally,therearestillseveralissuesthatneed
to be resolved. The biggest challenges are the high
frequent maintenance related to the power and
propulsion systems for large cargo ships and the
international regulations forʺfreeʺ sailing between
arbitrary ports. For a
small or medium sized ship
operating short distances within national waters,
these challenges can be overcome by using electric
propulsion with less maintenance needed, and by
obtaining national permits forunmanned operations
inlimitedareas.However,foracargoshipoperating
internationally,it ispossible toautomatesome tasks
for
certain operational phases within the next five
years, for instance mooring, cargo handling (cargo
dependent), and navigation in deep sea. Further
research is needed on maintenance free propulsion
systemsanddataandinformationneedsofanROC.
Moreover, on the regulatory side, it is necessary to
reviewtheCOLREGsandexamine
theimplicationsof
theMASScodeunderdevelopment.
ACKNOWLEDGEMENTS
The research leading to this paper is based on work
performed in several research projects including SFI
Autoship[40]fundedbytheResearchCouncilofNorway),
Autoship H2020 [41] funded by the European Unionʹs
Horizon 2020 program, and Smartare Transport [42] in
Møre & Romsdal county authority funded by the
NorwegianMinistryofTransport.
REFERENCES
[1]IMO, “IMO’s workto cut GHG emissions fromships,”
2024.
https://www.imo.org/en/MediaCentre/HotTopics/Pages/
Cutting‐GHG‐emissions.aspx(accessedFeb.28,2023).
403
[2]M. Kim, T.‐H. Joung, B. Jeong, and H.‐S. Park,
“Autonomous shipping and its impact on regulations,
technologies, and industries,” Journal of International
Maritime Safety, Environmental Affairs, and Shipping,
vol. 4, no. 2, pp. 17–25, 2020, doi:
10.1080/25725084.2020.1779427.
[3]A. Komianos, “The autonomous shipping era.
operational, regulatory, and quality
challenges,”
TransNav: International Journal on Marine Navigation
andSafetyofSeaTransportation,vol.12,no.2,2018.
[4]C.Barrera,I.Armas,F.Luis,O.Llinas,andN.Marichal,
“TrendsandChallengesinUnmannedSurfaceVehicles
(USV): From Survey to Shipping,” TransNav, the
International Journal on Marine Navigation and Safety
of Sea Transportation, vol. 15, pp. 135–142, Feb. 2021,
doi:10.12716/1001.15.01.13.
[5]H. Ringbom, “Regulating Autonomous Ships—
Concepts, Challenges and Precedents,” Ocean
Development&InternationalLaw,vol.50,no.2–3,pp.
141–169,2019,doi:10.1080/00908320.2019.1582593.
[6]Ø. J. Rødseth, “From concept to reality: Unmanned
merchant ship research in Norway,” in 2017
IEEE
Underwater Technology (UT), 2017, pp. 1–10. doi:
10.1109/UT.2017.7890328.
[7]C. Kooij and R. Hekkenberg, “Identification of a task‐
basedimplementation pathfor unmannedautonomous
ships,” Maritime Policy & Management, vol. 49, no. 7,
pp.954–970,2022,doi:10.1080/03088839.2021.1914878.
[8]C. Kooij, A. Colling, and C. Benson, “When Will
Autonomous Ships
Arrive? A Technology Forecasting
Perspective,” in A Technology Forecasting Perspective
(June6,2018).INECConference,2018.
[9]E. A. Holte, A. Pobitzer, H. Borgen, and Y. Chu,
“Smartere Transport – Møre og Romsdal A1.1
Ståstedsanalyse,”2019.Accessed:Feb.09,2023.[Online].
Available:
https://mrfylke.no/media/filer/samferdsel/vegavdeling/i
nfrastruktur/staastedsanalyse‐smartare‐transport‐sintef
[10]Ø. Rødseth, L.
A. Wennersberg, and H. Nordahl,
“Levels of autonomy for ships,” J Phys Conf Ser, vol.
2311, p. 12018, Feb. 2022, doi: 10.1088/1742‐
6596/2311/1/012018.
[11]O. E. Mørkrid, P. R. Bellingmo, and E. Wille,
“FeasibilityStudyforanUnmannedDeepSeaBulkShip
andShortSeaContainerShip,”2023.Accessed:Feb.09,
2023. [Online]. Available:
https://www.ntnu.edu/documents/1294735132/0/Feasibil
ity+study+UC1+and+UC2+unmanned+cargo+bulk+ship+
‐+signed.pdf/87eb1b15‐991c‐addf‐0a0c‐
319d1b547c19?t=1675860684177
[12]Sea Machines, “Collision & obstacle avoidance,” 2022.
https://sea‐machines.com/collision‐avoidance/ (accessed
Aug.22,2022).
[13]Avikus, “HiNAS‐Hyunday intelligent Navigation
Assistant System,” 2022.
https://www.avikus.ai/eng/product/hinas (accessed Sep.
05,2022).
[14]OrcaAI,“EnablingtheSmartShippingFuture,Today,”
2022.https://www.orca‐
ai.io/(accessedSep.05,2022).
[15]NYK Line, “Documentary of Fully Autonomous Ship
Project Released,” 2022.
https://www.nyk.com/english/news/2022/20220425_01.ht
ml(accessedFeb.03,2023).
[16]Kongsberg Maritime, “Situational Awareness,” 2022.
https://www.kongsberg.com/no/maritime/products/situ
ational‐awareness/(accessedFeb.03,2023).
[17]CaptainAI,“CaptainAI’sAutopilotsuccessfullytested
in the Port of Rotterdam,” 2020.
https://www.captainai.com/news/captain‐ais‐
autopilot‐
successfully‐tested‐in‐the‐port‐of‐rotterdam/ (accessed
Feb.03,2023).
[18]Hyundai Heavy Industries Group, “HD Hyundai’s
Avikus successfully conducts the world’s first
transoceanicvoyageofalargemerchantshiprelyingon
autonomous navigation technologies,” 2022.
https://www.prnewswire.com/news‐releases/hd‐
hyundais‐avikus‐successfully‐conducts‐the‐worlds‐first‐
transoceanic‐voyage‐of‐a
‐large‐merchant‐ship‐relying‐
on‐autonomous‐navigation‐technologies‐301559937.html
(accessedFeb.03,2023).
[19]Kongsberg Maritime, “Autonomous ship project, key
facts about YARA Birkeland,” 2023.
https://www.kongsberg.com/no/maritime/support/them
es/autonomous‐ship‐project‐key‐facts‐about‐yara‐
birkeland/(accessedFeb.03,2023).
[20]Ø. Rødseth, L. A. Wennersberg, and H. Nordahl,
“Improvingsafetyof
interactionsbetweenconventional
andautonomousships,”Feb.2021.
[21]International Maritime Organization, “Pilotage,” 2022.
https://www.imo.org/en/OurWork/Safety/Pages/Pilotage
.aspx(accessedMay09,2022).
[22]T. Porathe, “Where Does the Pilot Go When the
Autonomous Ship Has No Bridge? MASS Routing
Service and Smart Local Information Centres,” RINA
Autonomousshipconference,2022.
[23]MAiDSystems,
“MAIDSystems‐AFullyAutonomous
Docking and Positioning System for Recreational and
Commercial Marine Vessels,” 2022.
https://maidsystems.com/(accessedApr.27,2022).
[24]NYKLine,“NYKGroupCompanies ParticipateinTrial
toSimulatetheActualOperationofFullyAutonomous
Ship | NYK Line,” 2022.
https://www.nyk.com/english/news/2022/20220303_02.ht
ml(accessedApr.27,2022).
[25]Kongsberg
Maritime, “Ferry solutions,” 2022.
https://www.kongsberg.com/no/maritime/solutions/pax/
ferry/(accessedOct.06,2022).
[26]Nedcon Martime, “Crewstructure on board merchant
vessels‐engine department,” 2013.
https://nedcon.ro/crew‐structure‐board‐merchant‐
vessels‐engine‐department/(accessedMay03,2022).
[27]M.Hoyhtya,J.Huusko,M.Kiviranta,K.Solberg,andJ.
Rokka, “Connectivity for autonomous ships:
Architecture, use
cases, and research challenges,” in
International Conference on Information and
Communication Technology Convergence: ICT
Convergence Technologies Leading the Fourth
Industrial Revolution, ICTC2017, Dec. 2017, vol. 2017‐
December,pp.345–350.doi:10.1109/ICTC.2017.8191000.
[28]Ground Control, “Iridium Coverage Map‐Real Time
Tracking‐Ground Control,” 2023.
https://www.groundcontrol.com/en/knowledge/calculat
ors‐and‐maps/iridium‐coverage‐map‐
real‐time‐tracking/
(accessedFeb.06,2023).
[29]P.R.BellingmoandU.Jørgensen,“AutomaticMooring:
TechnicalGapAnalysis,”2022.Accessed:Feb.09,2023.
[Online]. Available:
https://www.ntnu.edu/documents/1294735132/0/sfi_auto
ship_automatic_mooring_technical_gap_analysis+%283
%29.pdf/71c86707‐ee5a‐2bae‐0fb7‐
a503f0aed1ef?t=1675872594993
[30]AutoMooringSolutions,“Automooringsolutions–AMS
AutoMooring Solutions,” 2023.
https://automooringsolutions.com/ (accessed Feb. 06,
2023).
[31]Trelleborg, “AutoMoor‐Automated Mooring
Solution,” 2023. https://www.trelleborg.com/en/marine‐
and‐infrastructure/products‐solutions‐and‐
services/marine/docking‐and‐mooring/automated‐
mooring‐systems/automoor(accessedFeb.06,2023).
[32]Cavotec, “Automated Mooring,” 2023.
https://www.cavotec.com/en/your‐applications/ports‐
maritime/automated‐mooring(accessedFeb.06,2023).
[33]MacGregor, “Automated mooring system,” 2023.
https://www.macgregor.com/intelligent‐
solutions/automated‐mooring‐system/(accessedFeb.06,
2023).
[34]Cavotec,“MoorMasterBrochure,”2021.
[35]iSAM
AG,“AutonomousGrabShipUnloadersforBulk
Materials,” 2023. https://www.isam‐ag.com/our‐
404
solutions/automation‐of‐grab‐ship‐unloaders‐for‐bulk‐
materials/(accessedFeb.09,2023).
[36]Norwegian Maritime Authority, “Rundskriv‐Serie V.
RSV12‐2020,”2020.
[37]IMO, “Autonomous shipping,” 2015.
https://www.imo.org/en/MediaCentre/HotTopics/Pages/
Autonomous‐shipping.aspx(accessedMar.01,2023).
[38]DNV, “Autonomous and remotely operated ships
(DNVGL‐CG‐0264),”no.September,2021.
[39]IMO,“Outcome
oftheRegulatoryScopingExercisefor
theUseofMaritimeAutonomousSurfaceShips(Mass),”
MSC.1/Circ.1638,vol.44,no.0,pp.1–105,2021.
[40]The Research Council of Norway, “SFI AutoShip:
AutonomousShipsforSafeandSustainableOperations‐
Prosjektbanken,” 2023.
https://prosjektbanken.forskningsradet.no/en/project/FO
RISS/309230?Kilde=FORISS&distribution=Ar&chart=bar
&calcType=funding&Sprak=no&sortBy=score&sortOrde
r=desc&resultCount=30&offset=0&Fritekst=sfi+autoship
(accessedJan.23,2023).
[41]Autoship, “Autoship,” 2023. https://www.autoship‐
project.eu/(accessedFeb.16,2023).
[42]Møre og Romsdal fylkeskommune, “Smartare
transport,” 2023. https://mrfylke.no/om‐oss/prosjekta‐
vaare/smartare‐transport(accessedFeb.16,2023).
