343

1 INTRODUCTION

Computational support for manoeuvers

determination in ship encounter situations may

significantlyincreasesafetyofpassengers, crewand

cargo. Various methods supporting collision

avoidance manoeuvers for ships have been

developed in the last decades. However, it was the

digitalandcomputersciencefastprogressinthelast

severalyearstha

tmadepossiblecreationofeffective

telematics systems implementing these methods.

Amonghardwaresolutionsonecandistinguishsuch

elements,alreadyimplementedonmostofvessels,as

AutomaticIdentificationSystem‐AISorTargetTracking

‐TT(includingAutomaticRadarPlottingAid‐ARPA).

These systems are utilized mostly to detect and

evaluate presence and motion parameters of the

othershipsinthevicinityoftheownship.

Software collision av

oidance solutions may be

roughly grouped into two categories: deterministic

and heuristic (or meta‐heuristic) methods. The

former one consists of algorithms that iteratively

calculate optimal and safe trajectories (or their sets)

in case of an encounter. A typi

cal property of a

deterministic collision avoidance method is a

possibility to obtain a fully optimal solution.

However, typically it is reached by investing

significant amount of processing and/or

computations, which results in long algorithm

execution time.Another disadvantage here is a

difficulty with handling some of the op

timization

constraints(e.g.dynamicconstraints).

Unlike the deterministic ones, the heuristic

collisionavoidancemethodsbaseonfastconverging

optimizationalgorithms,thususuallytheyareableto

return results after significantly shorter period of

time. Also possibilities to model virtually any

constraintintheoptimizationproblemareunlimited.

Yet, the ma

in disadvantage here is that the results

(trajectories or sets of trajectories) are usually sub‐

optimal.However,withafineconstructedheuristicit

is possible to obtain both fast convergence and

Data Acquisition in a Manoeuver Auto-negotiation

System

.Szlapczynska

GdyniaMaritimeUniversity,Gdynia,Poland

ABSTRACT:Typicalapproachtocollision avoidance systemswithartificial intelligence support is that such

systemsassumeacentralcommunicationandmanagementpoint(suchase.g.VTSstation),usuallylocatedon

shore. This approach is, however, not applicable in case of an open water encounter. Thus, recently a new

pproach towards collision avoidance has been proposed, assuming that all ships in the encounter, either

restrictedoropenwater,communicatewitheachotherandnegotiatetheirmaneuvers,withoutinvolvingany

outer management or communication center. Usually the negotiation process is driven by the collision

avoidancesoftwareandcalledauto‐negotiation.Thispa

perelaboratesondataacquisitionproblemincaseof

themaneuverauto‐negotiation.Itfocusesonshipsʹinitializationinthesystemanddatagathering.

http://www.transnav.eu

the International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 9

Number 3

September 2015

DOI:10.12716/1001.09.03.06

344

insignificant differences between optimal and sub‐

optimal solutions. One can distinguish also a

subgroup of heuristic collision avoidance methods,

namelytheexpertordecisionsupportsystems.These

systems usually relay on heuristics as their

optimizationcores,butextendtheirfunctionalityby

providingadditionaldecisionmakingsupport.

All the above mentioned

 approaches to

manoeuvers determination are based on an

assumption that the methods are used, utilized or

runinacentralcontroltrafficpoint,e.g.VTSstation,

usuallylocated on shore. Hence, a decision on how

the final manoeuvers should be like is made in an

arbitralandcentralizedfashion.While

thisapproach

is natural for restricted harbor area, hardly it is

possible for an encounter in open waters. Thus,

recently a new idea of decentralized manoeuvers

determination has emerged, as presented in (Hu et

al.,2008;Hornauer,2013;HornauerandHahn,2013).

It assumes that encountering ships communicate

with each

other without participation of any outer

party and establish their further collision avoidance

actionsbymeansofnegotiation.Ifthewholeprocess

of manoeuvers determination (i.e. data acquisition,

negotiation and optimization of manoeuvers) does

notneedanyhumaninteraction‐itiscalledanauto‐

negotiation one. Computational part of manoeuver

determination

 in the auto‐negotiation proposals

strongly relies on previously described centralized

approaches to collision avoidance, namely

deterministicorheuristicones(especiallythelatter).

However, in such case a few brand new elements

must be introduced to the system, such as detailed

data acquisition formats, negotiation protocols,

resultsdistribution,etc.

It

isworthnoticingthatthedecentralizedcollision

avoidance actions, often based also on negotiations,

are common for nowadays bridge communication

routines on open waters. Indeed, in a case of an

encounter,captainsornavigatorsoftheshipsusually

trytocontacteachotherdirectlyviaVHSradio(open

water) or via

 the VTS station (harbor area) to

establish necessary maneuvering. Obviously, in all

cases the arrangements ought to comply with

COLREGS(1972)regulations.

This paper aims at presenting the author’s

proposalofamaneuveringauto‐negotiationsystem,

particularly focusing on its first element: data

acquisition. The rest of the paper is

organized as

follows: section 2 presents a review of papers on

collision avoidance in general and also the auto‐

negotiation one. The section 3 presents a general

description of the proposed system. In the next

sectiondataacquisitioninthesystemispresentedin

detail (separately as initialization and data

gathering).

Section 5concludesand summarizesthe

materialpresented.

2 LITERATUREREVIEW

The centralized collision avoidance methods

constituteabasisofthemanoeuverauto‐negotiation

ones, thus a literature review of the former is

essential in this paper. Following previously

presented classification of collision avoidance

methods with a centralized control point,

they may

belong to either deterministic or heuristic group of

methods. A general, but thorough, review of both

thesegroupswaspresentedbyStatherosetal.(2008).

Classical deterministic collision avoidance

methodsarebasedondifferentialcalculusutilizedto

solvetheoptimizationproblem,aspresentede.g.by

Lisowski (1985). In this

 approach a set of ships is

monitored and controlled during the main

algorithm‘s run. It is based on a game model –

strategiesofallthevesselsaretakenintoaccountas

they were “thinking players”. The differential

elements are responsiblehere for modelling of ship

dynamic. Asimilar, differential

calculus‐based,

researchhasbeenconductedandpresentedin(Zak,

2004). It proposed utilization of multiple complex

motion rules describing collision situation, which

was a special case of a more general definition of

controlled movement approach. Another

deterministic method, called Collision Thread

Parameters Area (CTPA), was proposed by Lenart

(1982). In

 CTPA a navigator is able to select a safe

combination of own course and speed utilizing a

coupled coordinate polar system (presenting both

speed and position of ships), taking into account

circle‐shaped ship domains. The CTPA method has

been extended (FCTPA) to support any convex

domainbySzlapczynski(2008).

Mostoftheheuristiccollisionavoidancemethods

benefit either from genetic algorithms(GA) or their

successors – evolutionary algorithms (EA). One of

the first such approaches has been proposed by

Smierzchalski(1999),whereEAwasabouttofindthe

ownshiptrajectoryassumingthatalltheotherships

keeptheircourses

unchanged.Thisresearchhasbeen

continued since then, i.e. in (Kolendo et al., 2011) a

new scaling function has been proposed.Similar

collisionavoidanceGA/EAbasedmethodshavebeen

also proposed by other authors. Ito et al.(1999)

presented a GA algorithm utilized for collision

avoidancemaneuveroptimization.AlsoZeng

(2003)

and Tam & Bucknall (2010) proposed algorithms in

which optimal own ship trajectory is sought by

meansofanEAalgorithm.Anextendedapproach,in

which not only a single trajectory, but a set of

trajectories for all the ships in the encounter is

sought, has been proposed in

(Szlapczynski, 2010;

Szlapczynski&Szlapczynska,2012).Inthiscasenota

singletrajectory,butthewholetrajectorysetevolves

at once. Quite different, though still heuristic,

approach has been proposed byCheng& Liu (2007)

and Tsou & Hsueh (2010). The former presents

genetic annealing algorithm for trajectory

optimization, whereas the latter

 utilizes ant colony

algorithmforthesimilarpurposes.Anotherheuristic

ofswarmintelligencehasbeenutilizedinadecision

support system proposed by Lazarowska (2012). In

(Brcko and Swetak, 2013) fuzzy reasoning has been

usedtobuildacollisionavoidancedecisionsupport

system.Finally,Pietrzykowskietal.(2009)proposed

a

sophisticated decision aid system supporting

maneuveringcomplyingwithCOLREGSregulations.

Thefirstpapersonmanoeuversauto‐negotiations,

rejectinganaxiomofhavingacentralcontrol point,

emerged just recently. In papers (Hornauer, 2013;

HornauerandHahn,2013)adistributedsystemhas

been proposed in which ships negotiate their

345

manoeuvers. The authors plan to base the

optimization core on evolutionary sets of safe

trajectories,followingthepaper(Szlapczynski,2010).

Their proposal assumes that there is a distinction

between active ships, actively contributing to the

negotiation process, and inactive (rouge) ships. The

latter for various reasons (e.g. no proper

hardware/software

 equipment), do not assist in the

processanddo notfollowsystem recommendations

towards planned maneuverings. The system takes

into account presence of the inactive ships by

predicting (based on historic data) their future

movements.

Quite different assumptions lay at the basis of

CANFO system proposed by Hu et al.

(2008). Here

auto‐negotiation is only between two ships and

optimization core is reduced to a deterministic

method of manoeuver determination. The main

advantage of the proposal is a detailed negotiation

protocoldescription,basedonpre‐definedpreference

set. On the other hand, utilized straightforward

calculusmakesthemethodnotapplicable

incaseof

an encounter with more than two ships.What is

more, the authors assume that ship navigators

sometimeswouldrathergoforanillegal(according

toCOLREGS)maneuvering, if only economicfactor

(e.g. a way loss) would justify that. Such approach

hardlyisacceptableinpractice.

This paper

 aims at presenting the author’s

proposal of a manoeuver auto‐negotiation system,

deriving the advantages of the already known

solutionsandtryingtoavoidtheirdisadvantagesas

well as the known pitfalls. It is focused mainly on

ships’initializationinthesystemanddatagathering.

Itisfollowingauthor’sdraft

proposalof thesystem

presentedin(Szlapczynska,2014).

3 GENERALDESCRIPTIONOFMANOEUVER

AUTO‐NEGOTIATIONSYSTEM

Themainideaoftheproposedsystemistocombinea

manoeuvercontrol thatis independentofany outer

objects (such as e.g. a VTS station) with a semi‐

distributed control by the ships in

 the encounter.

Unlike previous manoeuver auto‐negotiation

proposals,theonecallsforcontrolflowinwhichone

oftheshipsintheencounterisaleadershipandall

the others are ships‐participants. The leader is

responsible for gathering data, determination and

optimization of manoeuvers, finally distribution of

the

results.Thispattern,calledheresemi‐distributed

one,asopposedtoadistributedone(inwhichallthe

shipsequallytrytocontributetothefinalresult),is

about to overcome the well‐known problems of

distributedsystem,suchasaccessingsharedmemory

ordeadlocks.



Figure1.Proposedmanoeuverauto‐negotiationsystem

Flow of the proposed system, presented in

Figure1, reflects the previously presented one in

(Szłapczynska, 2014). The following elements of the

systemhavebeendistinguishedthere:

1 Initialization–atthisstageshipscomeforwardto

beincludedintheauto‐negotiation.Theshipthat

initializes the process becomes

 a leader, all the

othersbecomeships‐participants.

2 Datagathering–alltheshipsreporttotheleader

theirbasicmovementparameterssuchascurrent

speed & course, initial and goal positions, etc.

Moreover,eachshipwilltransmitaninitialsetof

possible / forbidden courses (or possibly courses

& speeds). Both initializationand data gathering

constitute data acquisition stage, which is

elaboratedinthenextsection.

3 Manoeuver negotiation – here preferences

towards manoeuver possibilities and conditions

(e.g. circulation radiuses) are exchanged. A

detailedprotocoldescribingthenegotiationsisto

beprepared.

4 Trajectory determination and optimization –

collected

 input data (based on original plans

reported by all the ships, including also ships’

preferences towards maneuvering possibilities

and conditions) trigger the optimization process.

Itsearchesforanoptimal(orsub‐optimalrather)

setoftrajectoriesforalltheships.Itispossibleto

applyhereamultiobjectivesearchpattern,

thusa

search for multiple goals is possible (e.g. the

smallest way loss and the highest security of

traveling).Oneofthekeyconstraintshereisthat

optimization procedure must be executed on a

stricttimeregime,possiblyinlessthan1minute.

That forces utilization of heuristic approach to

optimization.

 Further research on a heuristic

methodsuitedfortheproblemisrequired.

5 Resultsdistribution–thelaststageintheprocess,

inwhichfinal resultsof computations(thesetof

trajectories)isdistributedamongtheshipsinthe

encounter. As the data is sent (via the selected

communication link)

as “suggested

maneuvering”, the system expects replies from

each ship (sent manually by the navigator)

confirming agreement for following the

suggestions.Onceallconfirmationsaregathered,

the results obtain a new status of “accepted

maneuvering”. The “suggested” status remains

unchanged in all other cases. No automatic

346

executionoftheacceptedmanoeuversisplanned,

it will remain in sole responsibility of the

navigators.Adetailedprotocoldescribingresults

distributionistobeprepared.

The system would be able to work in various

modes,dependingonthesituationandgeographical

localization of the encounter. These modes would

include

thefollowing(asnon‐excludingitems):

 openorrestrictedwatersmode,

 waters including / not including TSS (Traffic

SeparationSchemes)regulations;incaseofaTSS

Rule10will betakenintoaccount inmanoeuver

determination&optimization,

 forrestrictedwatersonly:watersunderVTS/no

VTScontrol,

 for good visibility (Rules 11 – 18 taken into

account)/restrictedvisibility(Rule19takeninto

account).

4 DATAACQUISITIONINTHESYSTEM

The first issue to be resolved when discussing data

acquisition in the system is which communication

channelshouldbeestablishedtoassurefast,reliable



and secure data exchange. As presented in

(Szlapczynska, 2014) there are mostly two

possibilities:

 AIS‐basedcommunicationbyutilizationofASM

messagebroadcast,

 wireless communication (“Wifi on sea” working

inpeer‐to‐peermode).

The key advantage of having the AIS‐based

channel is that its equipment, due

to international

regulations,isalreadypresentonmostofvessels(but

e.g. small vessels like fishery boats are still not

obligedtohaveAISonboard).However,thiskindof

communication relies on broadcasting messages

throughanunsecuredchannel.Thisfact,asreported

by(WWW_AIS,2015),maycausethefollowing:



fakeAISsignals(fromnon‐existingvessels)could

be transmitted, mostly to interrupt or collapse

systemsgatheringAISdata,

 AIS signals from existing vessels could be

amended e.g. by sending modified ship position

and/orcourse, which couldresultin initiationof

impropercollisionavoidanceactionsbytheother

vessels.

Thusallmayleadtoashipcollisione.g.

provokedonpurposebypirates.

Having the above in mind one should be aware

that no classified information should be exchanged

through the AIS channel, as it is defined at the

moment(thereareplans,however,torewritetheAIS

specificationand

thusinfuturetoincreasesafetyof

the connections). The problem with exchanging

classified information remains valid also for the

purposeoftheauto‐negotiationsystem.Ontheother

hand,thewirelessapproachwithTCP/IParchitecture

appliedisable to handle secure data exchange (e.g.

viadataencryption).Itsdisadvantage,

however,isa

necessity of having additional, quite expensive and

notwidespread,hardwareequipment. Thusthere is

no possibility to use this way of communication to

indicateship’swilltojointhesystem.Thereforeitis

recommended to initialize the auto‐negotiation

system via the AIS channel, but continue further

communication

 by establishing a secure wireless

peer‐to‐peerconnection.

The data acquisition element of the manoeuver

auto‐negotiation system may be split into two

separateitems,namely:

 initialization, when ships in an encounter that

would like to utilize the system set up a

negotiationgroup,

 datagatheringfrom

allships inthegroup, to be

used further by the trajectory determination &

optimizationcoreofthesystem.

The following subsections describe the

abovementionedelementsofthesystem.

4.1 Initialization

The main goal here is to set up a group of ships

interested in participation in the auto‐negotiation

system

andto select amongthema group’s leading

ship(aleader,theshipthatlaterwouldperformmost

ofthecomputationaltasks).Variousapproachesexist

tohandlesuchsituation.Asproposedpreviouslyby

theauthorin (Szlapczynska, 2014), the initialization

process (referred to as registration) would be

organizedasfollows.

Eachship,notinanencounter

currently, periodically sends an AIS message

broadcasting her will to become a leader of

maneuver negotiation. If two or more such ships

meet on a fixed area, e.g. an arena with 10Nm

diameter,thefirstonethatgetsintotheareabecomes

the leader and

 all the other become ships‐

participants. This approach in practice, however, is

limitedtofixedareas,e.g.precautionaryareas. One

also should have in mind that any COLREGS‐

compliant maneuver negotiation would be useful

only for more than a simple ship‐to‐ship encounter

(for two ships COLREGS state clearly

what is the

required maneuvering). Moreover, in some cases of

multiple ship encounter the leader would have

significantly smaller computational powers

(necessarytoperformmanoeuverdeterminationand

optimization)thantheotherparticipants,whichmay

leadtounacceptablelongresults’awaitingtime.

Toovercometheseproblemsadifferentapproach

toinitializationis

proposedhere.Lettheshipthatis

the most interested in manoeuver negotiation (i.e.

havingthebiggestnumberofconstraintsorshipsto

whom she is obliged to give way) to initialize the

negotiationbysendingtheAISmessagejustonce(or

eventuallyrepeatedbylimitednumberoftimes).The

process of sending the message could be triggered

eithermanuallybythenavigatororautomaticallyby

the system in case the Target Tracking (i.e. ARPA)

reports e.g. more than two other ships in possible

encounter with the own ship. All the ships (in AIS

range)thatreceivethemessage,thatwould

havethe

system installed and willing to participate, would

send back information “I’m in”, becoming

automatically ships‐participants. The others‐that

either would send “I’m NOT in” (having no such

system or not willing to participate) or not sending

any message‐would not become participants.

However,theirpresencewouldbe

noted(aspassive

ships) and monitored by the leader, further taken

intoaccountinallfuturecomputations.

347

4.2 Datagathering

Afterfinalizationoftheinitializationprocess,whena

group of ships‐participants, including a leader, is

established, every communication should be

performedonlyviaasecureconnection.Asdescribed

earlieritcouldbeachievedbyutilizationofwireless

peer‐to‐peer communication. As for now it is



assumed that such connection is available and

alreadysetupforthegroupofnegotiatingships.

In the next step all the participants should send

the leader information including their motion

parameters(e.g.currentcourse,speed,etc.).Basedon

theinformation,laterontheleaderwouldbeableto

determine

required and optimized maneuvering.

Obviously,thepassiveshipsbeing invicinityofthe

encounter would not send any information to the

leader.Thus, introducinga mechanismto overcome

this difficulty is necessary to obtain safe resulting

collisionavoidanceactions.

The author thus proposes that the participants,

instead of pure motion parameters,

 would send to

the leader information packs, so‐called here

“Manoeuver Availability Arrays” (MAA). A MAA

sent by a participant would consist of ranges of

courses(inanextendedversion:coursesandspeeds)

that would not result in a collision situation

(according to COLREGS rules) with any of the

nearby

ships.Thatwouldalsoincludeallthepassive

ships being nearby the participant, as well as any

other ship being out of the negotiation scope (not

visiblebythesystem,thusoutsideit)‐notactivenor

passiveonetothesystem.Thiswayeachactiveship

would introduce to the system

 additional

information about collision threats in the near

vicinity.TheMAAs could be determined (inalmost

real‐time fashion) in its basic form by utilization of

DCPA & TCPA values or in extended version by

utilizationofFCTPAmethod.

An example of MAAs utilization (in its basic

form)ispresented

inFigure2.Therearesevenships,

but only three of them are active ones in the auto‐

negotiationsystem(Ship1istheleader,Ships2&3

are participants). Within the range of the system

there is also Ship 4, but due to lack of proper

equipmentshestays

passiveanddonotcontributeto

thenegotiations(norsendsanyinformations).Ship5

is outside the scope of the system, but in near

proximitytothe Ship3.Similarly, Ship 6 is outside

butneartotheShip1andShip7isoutsidebutnear

totheShip

2.Thered&greencirclesaroundShips1‐

3 illustrate their MAAs determined by taking into

account all vessels nearby the considered ship. The

red MAA regions depict either forbidden (due to

COLREGSrules)orundesirednewcourses,whilethe

greenones– ranges of possible newcourses.Inthe



next process step (negotiation of maneuvers) the

greenandredareasmaybeslightlychanged,asthe

ships would agree on some tolerance of the own

undesired courses. Finally, the determination and

optimizationpartwilltakecareoffindingthebestset

ofmaneuveringforShips1‐3takinginto

accountthe

agreed(negotiated)possiblecourses(thegreenMAA

areas). Further research thus will focus on the

negotiation, optimization and results’ distribution

elementsoftheproposedauto‐negotiationsystem.



Figure2.AsamplecollisionsituationwithMAAsdepictedforactiveships(Ship1–3)inthemanoeuverauto‐negotiation

system

348

5 SUMMARY

Themanoeuverauto‐negotiationapproachoffersnot

only obvious improvement in automation of data

exchange between ships in an encounter situation,

but also may result in increased level of safety.

Primarily, the issue of human error in

communicationiseliminated in this case. It is often

that nowadays

 on international waters navigators

thatcommunicatewitheachother(eitherdirectlyor

via VTS) are not English native speakers. It may

result, especially in a stressful situation of an

encounter, that some language misunderstandings

would cause a collision. However, when designing

such a manoeuver auto‐negotiationsystem, one has

to

be also aware of COLREGS regulations. Only

systemsobeyingtheseregulationsareabletoassure

propergloballevelofsafety.

The proposal of a manoeuver auto‐negotiation

system presented in this paper is oriented at

designing flexible and robust procedures of

automatic ship communication while assuring

compliance with COLREGS. After completing

 the

nextstepsoftheresearch(designingthenegotiation,

optimizationanddistributionelementsofthesystem)

author aims at developing a system applicable in

practiceintothenavigationalworld.

REFERENCES

BrckoT.,SvetakJ.,FuzzyReasoningasaBaseforCollision

Avoidance Decision Support System. Promet – Traffic

&Transportation,Vol.25,No.6,str.555‐564,2013.

Cheng, X., Liu, Z., Trajectory Optimization for Ship

NavigationSafetyUsingGeneticAnnealingAlgorithm.

ICNC2007.Third InternationalConferenceonNatural

Computation.vol.

4,str.385–392,2007.

COLREGS,ConventionontheInternationalRegulationsfor

Preventing Collisions at Sea. International Maritime

Organization,1972(withamendmentsonDec2009).

HornauerS.,DecentralisedCollisionAvoidanceinaSemi‐

collaborative Multi‐agent System. Multiagent System

Technologies, Lecture Notes in Computer Science,

Volume8076,str.412‐

415,Springer,2013.

Hornauer S., Hahn A., Towards Marine Collision

Avoidance Based on Automatic Route Exchange.

Control Applications in Marine Systems, Volume 9,

Part1,str.103‐107,2013.

Hu Q., Yang C., Chen H., Xiao B., Planned Route Based

Negotiation for Collision Avoidance Between Vessels.

TransNav, the International Journal on

 Marine

Navigation and Safety of Sea Transportation, Vol. 2,

No.4,str.363‐368,GdyniaMaritimeUniversity,2008.

Ito,M.,Feifei Z., Yoshida,N., Collision avoidancecontrol

ofshipwithgeneticalgorithm.Proceedingsofthe1999

IEEEInternationalConferenceonControlApplications,

vol.2,str.1791–1796,1999.

Kolendo

P.,Śmierzchalski R., Jaworski B., Experimental

Research on Evolutionary Path Planning Algorithm

with Fitness Function Scaling for Collision Scenarios.

TransNav, the International Journal on Marine

Navigation and Safety of Sea Transportation, Vol. 5,

No.4,str.489‐495,GdyniaMaritimeUniversity,2011.

Lazarowska A., Decision support system for collision

avoidanceat

sea.PolishMaritimeResearch.Volume19,

IssueSpecial,str.19–24,DeGruyter,2012.

Lenart A.S. 1982. Collision threat parameters for a new

radar display and plot technique, The Journal of

Navigationvol.36:pp.404‐410.

Lisowski J., (1985). The analysis of differential game

modelsofsafeshipcontrolprocess.

JournalofShanghai

MaritimeInstitute,Volume6,No1,25‐38.

Pietrzykowski, Z.Magaj,J.Chomski, J., A navigational

decision support system for sea‐going ships. Pomiary,

Automatyka, Kontrola, R. 55, nr 10, str.860‐863,

WydawnictwoPAK,2009.

StatherosT.,HowellsG.,McDonaldMaierK.,Autonomous

Ship Collision Avoidance Navigation

 Concepts,

Technologies and Techniques. The Journal of

Navigation, 61(01), str. 129‐142, Cambridge University

Press,2008.

Smierzchalski,R.,Evolutionarytrajectoryplanningofships

in navigation traffic areas. Journal of Marine Science

andTechnology,vol.4,Issue1,str.1–6,Springer,1999.

Szlapczynska J.: Propozycja systemu auto‐negocjacji

manewrów statków korzystają

cego z metod

optymalizacji wielokryterialnej oraz Matematycznej

TeoriiEwidencji (inPolish), Logistyka vol. 6/2014, pp.

10375‐10384,2014.

Szlapczynski R.: Fuzzy Collision Threat Parameters Area

(FCTPA) – A New Display Proposal. TransNav, the

InternationalJournalonMarineNavigationandSafety

ofSeaTransportation,Vol.2,No.4,pp.359‐362,2008



Szlapczynski R., Solving Multi‐Ship Encounter Situations

by Evolutionary Sets of Cooperating Trajectories.

TransNav, the International Journal on Marine

Navigation and Safety of Sea Transportation, Vol. 4,

No.2,str.185‐190,GdyniaMaritimeUniversity,2010.

Szlapczynski R., Szlapczynska J., On Evolutionary

ComputinginMulti‐ShipTrajectoryPlanning.Applied

Intelligence,

Volume37,Issue2,str.155‐174,Springer,

2012.

Tam,C.,Bucknall,R.,Path‐planningalgorithmforshipsin

close‐range encounters. Journal of Marine Science and

Technology,vol.15,Issue4,str.395‐407,Springer,2010.

Tsou, M. C., Hsueh, C. K. The study of ship collision

avoidance route planning by

 ant colony algorithm.

Journalof Marine Science andTechnology, 18(5), 746–

756,Springer,2010.

Zeng X., Evolution of the safe path for ship navigation.

AppliedArtificialIntelligence.17,str.87–104,Taylor&

Francis,2003.

Zak B., The problems of collision avoidance at sea in the

formulationofcomplexmotionprinciples,Int.

J.Appl.

Math.Comput.Sci.,2004,Vol.14,No.4,pp.503–514.

WWW_AIS, http://www.aisreporter.com/?news=ais‐

vulnerabilities‐subject‐of‐scientific‐research, web page

accessedon2015.02.01,2015.

