391
1 INTRODUCTION
Intelligent,machine‐controlledtransportationsystems
are becoming increasingly popular due to their
increased efficiency relative to systems relying on
humanoperators.Theever‐increasingperformanceof
computersystems,aswellasthedevelopmentofnew
remotesensingsystems,coupledwiththefallingcosts
of implementing the latest solutions,
allows for the
continued expansion of automated transportation
systemsapplicationrange.
In the field of maritime transportation, there is
considerable interest in the issue of its automation,
bothamongresearchersandshipowners.Theseefforts
are centered around the idea of MASS (Maritime
Autonomous Surface Ships). MASS are expected to
use
arangeofdevelopingtechnologiesthatwillallow
them to move autonomously through the water and
perform tasks, first under constant operator
supervision and eventually under incidental
supervision.
The main advantage of the widespread
introductionofMASSwillbetheincreasedsafetyof
navigationcausedbytwofactors[1].Theabsence
ofa
crewonacommercialvesseleliminatestheriskofloss
of life in the event of a serious accident, including
sinking(excludingpassengervessels).Secondly,the
increasednumberofsensors,andtheircouplingwith
a powerful computer data processing system, can
allowamuchmorepreciseassessmentof
thesituation
bythealgorithmscontrollingtheunmannedvessel.
Another advantage of MASS will be the
optimization of the unitʹs operating costs [2].
Regardless of the propulsion system used:
conventional, hybrid (e.g., diesel‐electric) or all‐
electric,thecontrolalgorithmswillbeabletotakeinto
accountthecurrentoperating
pointofthepropulsion
systematanytime duringthe voyage. Knowingthe
Control of Electric Drive Tugboat Autonomous
Formation
W.Koznowski&A.Łebkowski
GdyniaMaritimeUniversity,Gdynia,Poland
ABSTRACT: The automation of maritime transport is an indispensable trend towards full autonomy of
maritimevessels.Inthispaper,anattemptwasmadetopresentthecontrolsystemforportautonomousvessels
usinganagentsystem.Onthebasisoftheconductedresearch,inorderto
optimizetheenergyconsumption
relatedtothemovementoftugboats,theshapeofthehullandtheshapeoftheformationinwhich4tugboats
aremovingwereselected.Severalscenariosofnavigationalsituationsthatmaytakeplaceinportwatershave
beenrecognized.Theconductedanalysishaveshownthat
theoptimalshapeofthehulloftugboats,theshape
oftheformationinwhichtheymove,aswellasthedeterminationofthepassageroutefortheimplementation
of a specific task, can contribute to reducing both the carbon footprint and the energy consumption of the
propulsionsystemsoftugboats.
Thisisofsignificantimportanceintermsofreducingexhaustgasemissionsin
andaroundports.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 17
Number 2
June 2023
DOI:10.12716/1001.17.02.16
392
performancecharacteristicsof thepropulsionsystem
allowstheshiptoadjustthesailingconditionsfound
by the vessel, such as existing and forecast
hydrometeorological conditions, current navigation
conditions,oranticipatedwaitingperiodsforservice
inport.
Iftherearisesaneedtospeedupthevoyage,the
autonomous system will
also be able to take into
accounttheaforementionedconstraintsandcarryout
theappropriateoptimizationofitspassageroute,so,
for example, it hits the previously released time
windowforimmediateportentry,rightontime.
Anaspectthatiscloselylinkedtooperatingcosts
isenergyconsumptionand
theassociatedamountof
pollution emitted into the environment. There is no
doubtthatproperplanningoftheroute,andaquick
and appropriate response to changing sailing
conditionshaveasignificantimpactontheamountof
fuelrequiredtocoveragivendistance.
PracticalexamplesofMASSunitscurrentlyexist
in
the prototype phase and in some cases there are
alreadyoffered as products. These are usually small
ships,equippedwithmanydifferenttypesofsensors
and computer systems using their signals. An
exampleofsuchaprototypevesselistheMayflower
ship [3]. It is a 15‐meter trimaran,
equipped in
addition to classic navigation devices such as AIS,
GPS,Electronic Chart, Echosounder,and Radar,also
with 6 cameras supported by artificial intelligence
(AI)technology.
Figure1.AutonomousshipMayflower[3].
Additional information on the condition of the
Mayflower is provided by sensors of the physical
position of the hull, an on‐board diagnostic system
providinginformationonthestatusofthepropulsion
and telecommunications systems, and a remote
weatherforecastservice.
AnotherexampleofaprototypeisAutoNaut[4],a
small
test vessel designed to study a propulsion
system that uses wave energy to move. It does not
have such a developed autonomy system as
Mayflower, but it uses radar and AIS for automatic
collision avoidance. An additional feature that
increasessafetyisthepaintingofthehullinabright
color,aimedatimprovingvisibility.Remotecontrolof
the unit and data reading was carried out using a
satellitelink.
Figure2.AutonomousvesselAutoNaut[4].
AnexampleofaMASSunitalreadyinproduction
istheC‐Worker7,offeredbyL3HarrisTechnologies.
Itisanoffshorevesselwithalengthof7.5m,awidth
of2.3mandamaximumspeedof6knots.C‐Worker7
isintendedmainlyforinspectiontasksor
asabasefor
positioningsystems.Itisequippedwithamoonpool
and has the ability to carry an underwater vehicle
(ROV)orvarioustypesoftranspondersofpositioning
and measurement systems, such as Ultra Short
Baselineacousticpositioningsystem(USBL),Acoustic
Doppler Current Profiler (ADCP), or multi‐beam
sonar.
This ship is equipped with a typical set of
classicnavigationsystems,including:radar,compass,
inertial navigation system (INS) and AIS. The
autonomoussystemisbasedon4camerasoperating
invisiblelightand6camerasoperatinginthethermal
infrared(IR)range.
Figure3.AutonomousshipC‐Worker7[5].
AutonomousshipsofferedbyL3Harris,including
C‐Worker7,usetheASviewsoftware,consistingof4
elements:thecore,integratingsignalsfromtheshipʹs
sensors and the autonomy system processor, radio
communication system, operator station, and
optional, portable remote control system. The
manufacturer declares that the autonomous
navigationsystemcomplies
withtheCOLREGrules.
TheASviewsoftwarealsoallowsvesselstooperate
in a formation mode limited to 4 units, mainly to
allowasingleoperatortosendseveralvesselstothe
sameareaofoperation.
The subject of this publication is to present an
outline of the structure and
capabilities of the agent
system designed to control a formation of electric
tugboats.Themainpurposeofthesystem,inaddition
tosafecontroloftheformationʹsrouteinaccordance
393
withtheadoptedCOLREGrules,willbetominimize
the consumption of energy intendedfor moving the
formation on the route leading to the destination
point.
Against thebackgroundof the existing literature,
there is currently no system that allows the VTS
operatortoautomaticallyperformvarious porttasks
using
a formation of autonomous tugboats. The
system proposed by the authors, will allow the
operatortoindicatethepurposeofthetask.Thenthe
systemwillmakeitpossibletoplanandcarryoutthe
routeofthetugsfromthedockingpointtotheplace
ofoperation,performthe
orderedtasksandreturnthe
tugstotheplaceofdocking.
2 ELECTRICTUGBOATFORMATION
Theproposedagentsystemwillbeusedtocontrola
formationconsistingofporttugboats,equippedwith
an electric propulsion system, using at least two
azimuth propellers to ensure adequate
maneuverability.Thesizeofthesupported
formation
startswithaminimumnumberoftwounits,withno
fixed maximum. However, the practical number of
tugboats used will depend on the type of tasks
performed,andfortypicalharbortugstasks,suchas
shipescort,itwillbebetween2and5tugs.
The basic task of
the tugboat formation control
system will be to carry out operations related to
securingtheshipsenteringtheport,leavingtheport,
andmovinginsidetheport. Thecontrolsystemwill
also enable the use of tugboats of the formation to
performothertypesoftasks,which,duetothehighly
economical propulsion system and the lack of crew,
couldbesuccessfullyperformedbytugboatsinplace
ofotherunitsusuallyusedforthispurpose.
An example of such tasks may be ice actions
performed in the waters of the port. In weather
conditions causing the waters of the port to
freeze,
unmannedtugboatsfreefromothertaskscanbesent
tobreaktheicelayer,whichwillnotrequiretheuseof
icebreakers. Another possible application of waiting
tugboats may be performing patrol functions,
combinedwithcontinuousmeasurementofthedepth
oftheportbasin.
Afterappropriateadaptationofthetugboat
itself,
itwouldbepossibletoperformadditionalfunctions,
e.g.: transporting the Pilot to a ship waiting in the
roadsteadorgettingthePilotbackfromashipleaving
theport,performingpollutionmonitoringoperations
[6], and cleaning the port waters from spills and
floating garbage, using hull‐mounted collecting
devices[7].
2.1 Electricpropulsionandrecharging
The energy to propel the tugsʹ propellers will come
fromtheon‐boardenergystoreusingelectrochemical
cells, recharged from an external charging station
locatedonthequay.Oneoftheimportanttasksofthe
agentsystemwillbetotakecareofthe
stateofcharge
ofthebatteriesofallvesselsintheformation.
This task includes following aspects: monitoring
theactualStateofCharge(SOC)oftheenergystorage,
monitoring differences in the State of Charge and
energy consumption between each of formation
vessels,monitoringtheavailabilityofvarioustypesof
energy sources available at a given moment, with a
particularfocusonrenewableenergysources.
InordertomonitortheStateofChargeofthestore
and to detect differences in the operation of stores
locatedondifferentvessels,thesystemcanusesetsof
sensors built into the energy stores.
Knowledge of
datafromtwinvesselswillenabletheuseofstatistical
methods to capture symptoms heralding the
imminentfailures.
On‐board energy store can be recharged via a
charging station located near the mooring place for
tugboats of the formation. In order to reduce the
impact on the natural
environment, the charging
stationcanuseenergyavailablefromlocalrenewable
sources,suchasphotovoltaic,wind,andwaveandsea
current power plants. The employment of modern
weatherforecastingsystemsmayenableplanningthe
charging process for periods when the appropriate
amountofenergyfromthesesourceswillbeavailable.
In
theeventofunfavorableconditionsforaccessto
renewableenergy,e.g.:duringnight,cloudyweather,
orwindlessweather,itmaybenecessarytocarryout
the charging process using energy from the grid,
which may involve the emission of CO2 and other
pollutants, proportionally to the grid load by the
chargingstation.Minimizingtheenergyconsumedby
tugboatsthereforehasadirectimpactontheamount
ofpollutantsemittedand theavailablemethods and
possibilitiesshouldbeusedtoreduceitsconsumption
inordertoachievethehighestlevelofsustainability
[8].
2.2 Multiagentcontrolsystem
Itisplannedtoemploy
anagentsystemtocontrolthe
formation of electric tugboats. According to the
definition[9],agentsareautonomousunitsthathave
the ability to operate in a specific environment.
Agents can communicate with other users of the
environment and use owned, or foreign resources.
Agentsmay have specifictasks,or
objectives, which
may be general, or specific. An environment is a
certainspacewhereagentsareplaced,wheretheycan
act, and influence this environment. Agent systems
areoneofthemanymethodsofartificialintelligence
imitating natural phenomena. Another well‐known
bio‐derived method of artificial intelligence are
ArtificialNeural
Networks(ANN)[10].
The operating environment of the agent system
controlling the formation of tugboats is the physical
space covering the waters of the port in which the
system operates, and the virtual space of computer
systems located in the port area and on board the
formationtugboats.Theenvironment
islocatedinthe
areaoflimitedwaters[11]andthereforetherequired
responsetimetoachangeinnavigationconditionsis
shorter than in open waters, while the possibility of
interferencewithnavigationalinstrumentsisgreater.
On board a single tugboat there is a computer
system that collects signals from
navigation devices
394
and sensors of the tugboatʹs devices [12]. The
navigationdevicesusedmayinclude:radar,GPS,AIS,
compass, Inertial Navigation System (INS), echo
sounder, and an electronic map [13] of the area in
which the formation operates. The sensors include,
among others: sensors of the propulsion system,
sensors of the
electric energy store, and optionally
remotesensingsensorsusedbythetugboattodetect
objectsinthevicinityofitshull.Itispossibletouse,in
particular,sensorssuchas:lidarandcamerasystems
operating in visible light, near infrared or thermal
infrared.Thetugboatmustalsobeequipped
withan
appropriate telecommunications infrastructure,
providingatleastabasicandbackupcommunication
channeltoothersystemparticipants.Thestructureof
the control system for an autonomous formation of
porttugsisshowninFig.4.
Figure4.Autonomousformationcontrolsystemstructure.
Thestructureofthe systemis hierarchical,where
thebasiclevelofthehierarchyisasingletug.There
are several agents in its on‐board computer system
whose tasks include: communication with the
environment and supervision over the work of on‐
board agents, observation of the environment using
availablesensors
and signalsfromother participants
intheenvironment,controlofthepropulsionsystem,
supervision of on‐board systems, in particular
electricitystorage.
Thesecondlevelofthehierarchyistheformation
itself,madeupofindividualtugs.Itscompositioncan
bedynamicwhen,forexample,itbecomesnecessary
toreplace
atugboatthathasfailed,orwhen, dueto
unforeseen deterioration of hydro‐meteorological
conditions, it becomes necessary to increase the
numberoftugsconstitutingaformation.
3 SIMULATIONRESEARCH
Onthebasisofthepreviouslymentionedassumptions
regarding the tasks possible to be performed by the
formation, tests were carried
out using a simulated
formationconsistingof4tugsoperatingintheportof
Gdańsk.Ascenariowasperformedwhereaformation
of4tugboatsleavesthebasin,whereitiswaitingfor
the assignment of tasks, in order to meet the ship
awaiting on the roadstead for assistance
in entering
theport.Thesituationalviewoftheanalyzedscenario
isshowninFig.5.Asaresultofpreviousresearch,it
wasdeterminedthat themostenergy‐efficientshape
of the formation of 4 tugboats is a linear formation
with a small gap between the tugboats flowing one
after
the other. Such arrangement of the tugboats
allows for the reduction of hydrodynamic resistance
byapprox.57.6%.Inturn,theuseofanelectricdrive
powered by an energy storage system instead of a
conventional diesel drive reduces energy
consumption by up to 75.23% compared to the
potentialenergycontained
indieseloil.
Figure5. View of the simulated port of Gdańsk with
superimposedplotofroutetraveledbytheformation.
In Fig. 5, the dotted blue line shows the route
traveled by the tugboatsof the formation controlled
bytheagentsystem.Eachdotrepresentstheposition
of the first tugboat of the formation approximately
every16softhesimulationtime.Aftertravelingabout
1km,theagentresponsiblefornavigation,
locatedin
thefirsttugboat(theformationleader),hasdetecteda
ship leaving the quay, accompanied by two other
tugboats. The agent decided to avoid the working
area of the foreign tugboats and changed the
formationroute,sailingaroundtheislandbreakwater
from the eastern side, and then continued sailing
northtowardsthetargetship.
Figures6÷9showgraphsofthedataobtainedasa
resultofthesimulation.
Figure6.Tugformationleaderspeedplot.
Fig. 6 shows the speed chart of the first tugboat
(leader)oftheformation.Thereareslightslowdowns
inplaceswherethecoursechanged.
395
Figure7.Plotofthecourseandangleofthethrustersofthe
firsttugoftheformation.
In Fig. 7, the green line shows the course of the
firsttugoftheformation,andthebluelineshowsthe
deflectionangleofthetugʹsazimuththrusters.
Figure8. Plot of wind speed and wind angle felt by the
leadingtugoftheformation.
In Fig. 8, the teal line shows the direction of the
wind,andtheorangelineshowsthewindspeed.Both
valuescorrespondtotheconditionsmeasuredbythe
firsttugoftheformation.
Figure9.Plotofthespeedandangleoftheseacurrentfelt
bytheleadingtugoftheformation.
InFig.9,thepurplelineshowsthedirectionofthe
sea current, and the gray line shows its speed. Both
valuescorrespondtotheconditionsmeasuredbythe
firsttugoftheformation.
Asaresultofsimulationtests,thefollowingresults
were obtained for a formation consisting of 4
tugs.
Distancecovered:7581m,energyconsumptionbythe
formation for electric drive: 214.9 kWh, respectively
diesel consumption of conventional propulsion 71.6
dm3 (761.8 kWh). CO2 emission for electric
propulsion assuming energy use from the public
network was 71.8 kg, and 189.1 kg using a
conventionaldrive.
For comparison, in
the case of 4 free‐flowing
tugboats,notmaintainingaparticularformation,the
electricenergyconsumptionwas366.6kWh,andthe
dieselconsumptionwas122.2dm3(1300.2kWh).The
CO2emissionfortheconsiderednavigationsituation
wasrespectively:122.4kgforelectricdriveand322.6
kgfordieselpropulsion.
Asa
resultofonlytheuseofelectricpropulsion,a
reduction of CO2 emissions of 62.0% was achieved,
and 77.7% as a result of combining the benefits of
electric propulsion and the effects of reducing
hydrodynamic resistance due to the movement of
tugboatsinalinearformation.
4 SUMMARY
The article
presents the concept of an agent system
designed to control an unmanned formation of
electric tugboats. The use of an agent system
embedded in the physical environment of the port
andthecomputersystemoftugboatsmadeitpossible
todefineasystemthatsignificantlyextendsthescope
of tasks previously
performed by tugboats. In
additiontotypicalassistanceoperations,idletugboats
will be able to perform diagnostic or cleaning tasks
without the need to engage additional vessels and
theircrews.
The research showed that a formation controlled
byanagentsystemcanreactinaccordancewiththe
COLREG rules to
unforeseen situations, such as
findingapairoftugboatsstartingtosteerashipout
oftheportwithintheworkingarea.
Theuseofanagentsystemallowsforcontinuous
monitoringofthenavigationsituationandreactingto
changing conditions that may increase energy
consumption. When there are several options
for
dealing with an urgent situation requiring a course
correction, the tug formationʹs team of agents can
analyzeeachofthepossiblecandidatemaneuversto
selectthemostenergy‐efficientone.
The application of an electric drive system also
reduces operating costs by reducing energy
consumption and reducing pollutant
and CO2
emissions. The conducted research showed a
reductioninenergyconsumptionbyapprox.72%by
the electric drive compared to the combustion drive
andareductionofpollutantemissionsatthelevelof
approx.78%.
REFERENCES
1.Vos, J. de; Hekkenberg, R.G.; Valdez Banda, O.A. The
Impact of Autonomous Ships on Safety at Sea – A
Statistical Analysis. Reliability Engineering & System
Safety2021,210,107558,doi:10.1016/j.ress.2021.107558.
2.Krell, E.; King, S.A.; Garcia Carrillo, L.R. Autonomous
SurfaceVehicleenergy‐efficient andreward‐basedpath
planning using Particle
Swarm Optimization and
Visibility Graphs. Applied Ocean Research 2022, 122,
103125,doi:10.1016/j.apor.2022.103125.
3.Mayflower Autonomous Ship. Available online:
https://mas400.com(accessedon20February2023).
396
4.AutoNaut, wave propelled uncrewed surface vessels.
Availableonline:https://www.autonautusv.com/vessels‐
0(accessedon22February2023).
5.L3Harris. C‐Worker 7 ASV. Available online:
https://www.l3harris.com/all‐capabilities/c‐worker‐7‐asv
(accessedon22February2023).
6.Veettil,B.K.;HongQuan,N.;Hauser,L.T.;DoanVan,D.;
Quang,N.X.Coastalandmarineplastic
littermonitoring
usingremotesensing:Areview.Estuarine,Coastaland
Shelf Science 2022, 279, 108160,
doi:10.1016/j.ecss.2022.108160.
7.Choe,Y.;Lee,J.;Jung,W.;Park,J.;Lee,J.; Jho,J.Y.;Lee,
K.T.; Kim, T.; Kim, Y.H. Gravity‐based oil spill
remediationusing reducedgrapheneoxide/LDPEsheet
for both light and heavy
oils. Process Safety and
Environmental Protection 2021, 156, 617–624,
doi:10.1016/j.psep.2021.10.045.
8.Przybyłowski,A.;Kałaska,A.;Przybyłowski,P.Questfor
a Tool Measuring Urban Quality of Life: ISO 37120
StandardSustainable DevelopmentIndicators.Energies
2022,15,2841,doi:10.3390/en15082841.
9.Rossi, F.; Bandyopadhyay, S.; Wolf, M.; Pavone, M.
Review of
Multi ‐Agent Algorithms for Collective
Behavior: a Structural Taxonomy. IFAC‐PapersOnLine
2018,51,112–117,doi:10.1016/j.ifacol.2018.07.097.
10.Staniek,M.; Czech,P.Self‐correctingneural networkin
roadpavementdiagnostics.AutomationinConstruction
2018,96,75–87,doi:10.1016/j.autcon.2018.09.001.
11.Kaizer, A.; Winiarska, M.; Formela, K.; Neumann, T.
Inland Navigation as an Opportunity to
Increase the
CargoCapacityoftheTri‐CitySeaports.Water2022,14,
2482,doi:10.3390/w14162482.
12.Buc, A.; Bielecka, A.; Iwaszkiewicz, J. The Use of
Wireless Networks and Mobile Applications in the
ControlofElectricMachines.AppliedSciences2023,13,
532,doi:10.3390/app13010532.
13.Weintrit, A. Time to Revise the IMO’s Guidance on
Good Practice for the Use of Electronic Chart Display
and Information System (ECDIS). TransNav 2022, 16,
523–531,doi:10.12716/1001.16.03.15.