International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 4

Number 4

December 2010

479

1 INSTRUCTION

The environment of navigation has great change in

recent years. This make the maneuvering of ship be

more difficult. At the same time, ARPA (Automatic

Radar Plotting Aids), GMDSS(Global Maritime Dis-

tress &Safety System), GPS(Global Positioning Sys-

tem) and ECDIS(Electronic Chart Display and in-

formation System) etc. have applied in navigation,

the number of crew is decreasing. This make more

serious for manipulator. Eight percents of shipwreck

accident were caused by human factor according to

investigation.( Guedes Soares & Teixeira 2001,

Gaarder et al. 1997) To decrease the accident, and

increase the safety of navigation, researchers bring

in automatic maneuvering to instead of human’s job.

This method makes up for some human’s shortage,

and increases work efficiency. However, they find

that some accidents related to automatic equipment

late years. Some person research the strand accident

of “Royal Majesty” find that automation changed the

style of working, and formed a new way to making

mistake. So this paper proposes an intelligent evalu-

ation system of ship maneuvering, human is the ma-

jor in system, the system can calculate the status of

ship and collision risk, then display it functionality.

If the vessel gets into the critical area and time, the

system will adopts corresponding strategy when the

navigator does not adopt any measures.

2 HUMAN FACTOR

According data of Japanese Ship Safe Seminar in

1998, 84% shipwreck accident caused by human

factor. Other country and region also have similar

conclusion. The human’s factor attaches importance

to navigation safety, which has turn into people’s

consensus. Human have researched into human’s

factor indefatigably for ages, and form an academic

domain “Human Factor”(Gaarder et al. 1997). With

the purpose of enhancing security and efficiency, it

is an extremely practical integrated subject. Human

factor can bring positive impact on traffic at sea, for

example human’s cognitive and perceive capacity is

stronger than instrument and equipment. It can bring

negative impact reversely, for example making a

mistake easily, forgetting memorial affair, limited

analysis precision. Human’s fault is one of major

reason in shipwreck, and comes in for human’s high-

ly respect.

Fault is the property of human, removing the fault

completely is an unpractical. Therefore, we should

adopt measures to reduce the harmful consequence

brought by man. To achieve a job, division of labor

is an efficacious practice. BRM (Bridge Resource

Management) and BTM (Bridge Team Manage-

ment) at sea are discussed more by researchers of

last year. VTS is a system of construct with vessels

taking part in VTS and VTS organization. It has cor-

rection capability on the part of whole system. VTS

attendant will correct it when ship has breach of reg-

Intelligent Evaluation System of Ship

Management

Q. Xu & X. Meng

Information Science and Technology College, Dalian Maritime University, Dalian,

China

N. Wang

Marine Engineering College, Dalian Maritime University, Dalian, China

ABSTRACT: The security of maritime traffic is a significant part of intelligent maritime traffic. It can reduce

to ship maneuvering and collision avoidance by macroscopic. Eighty percents of marine accident induce by

human factor from research data. So some researches about intelligent computer evaluation system to reduce

the accident of human caused have emerged. Intelligent evaluation system of ship maneuvering can calculate

the status of ship and getting the data of ship around, and then adopt fuzzy comprehensive evaluation method

to calculate the collision risk and evaluate the operation of navigator. If it has danger of collision risk or the

navigator adopts irrational operation scheme by calculating, the system will send message to the navigator.

The navigator must affirm the messages, if there is not affirmance, the system will adopt collision avoidance

measures or other rational operations automatically at the critical moment.

480

ulation phenomenon. Another method to reduce fault

is using alerting equipment. The equipment send out

warning to cause human’s attention when mistake

occurs. For example, ARPA can send out sound and

light warning when the DCPA and TCPA are small-

er than a setting value. Britain, Germany and Japan

develop BNWAS used to monitor steering and sail-

ing on duty. This system used to monitor the alert of

navigator, if the equipment detects that navigator

cannot perform the duty of his, it will send out gra-

dated outspread warning. At first, it will be in cage,

if there is no response it will extend to caption and

other sailor’s room.

3 INTELLIGENT EVALUATION SYSTEM

Traditional collision avoidance is that the sailor

adopts empirical collision avoidance according to

self-experience. It depends on navigator’s individual

intuition to make decision, if the risk is large, it will

be easy to make mistake. Collision avoidance expert

system and decision-making support system spring

up rapidly of late years. They have great auxiliary

effect to vessel collision avoidance.

Human is the principal part in the evaluation sys-

tem of ship maneuvering. We make use of computer

and develop intelligent evaluation system of ship

maneuvering. The system can gather dynamic in-

formation of vessel by AIS, ARPA, infrared and

photo electricity equipments(Thomas et al. 2008).

The information will be sent to the intelligent evalu-

ation system finally, the system will enter into dif-

ferent model according to encounter situation and

environment condition. The result of evaluation is

the current situation of ship.

3.1 The Structure of Intelligent Evaluation System

The evaluation system consist of many models, in-

cluding target ship identification, speculation and

prediction of encounter status, real evaluation of op-

eration, auto-collision avoidance strategy and risk

warning model etc. We can see from Fig. 1, the

evaluation system and operation of navigator form a

closed-loop control system. The system will evaluate

the performance of operation, and send out corre-

sponding signals. In this way it can make up the dis-

advantage of none precision calculation of human,

cut down the probability of human fault occurrence,

and secondly make use of human’s high adaptability

sufficiently.

3.2 Collision Risk Calculation

Ship collision risk calculation is one of the most im-

portant parts in the system. The quantification of

collision risk experience several stages basical-

ly(WU Zhao-lin & ZHENG Zhong-yi 2001). The

first one is traffic flow theory which use ship colli-

sion rate, encounter rate, collision probability to

evaluate the collision risk for special water area. The

second is ship domain and arena which is based on

human praxiology and psychology. (Fuji & Tanaka

1971), (Goodwin 1975) etc. who use this to calculate

collision risk. In the third stage, people have consid-

ered the dCPA(Distance at Closest Point of Ap-

proach) and tCPA(Time at Closest Point of Ap-

proach) in calculation, like (Davis et al. 1980). In the

fourth stage, combine dCPA and tCPA, adopt

weighting method to calculate collision risk at the

beginning(Kearon 1979, Imazu& koyama 1984).

This method exist obvious disadvantage that dCPA

and tCPA are two different variable. Then people

adopt fuzzy theory to combine dCPA and tCPA. At

present mostly research are based on the artificial in-

telligent technology as fuzzy theory, expert system,

neural network to calculate the collision risk(LI Li-

na 2006).

This paper adopt fuzzy compressive evaluation to

calculate CR(collision risk). The comprehensive

evaluation result can be used as subjective evalua-

tion, and also can be as objective one. Furthermore,

system security is a progressively process. We can

get perfect result through assessing the subordina-

tion of the factors. So we don’t use the weighting of

dCPA and tCPA to calculate collision risk, they ap-

plied fuzzy comprehensive evaluation in it. There

are many factors effecting CR. We only consider the

major factors here, the distance between target ship

and local ship d, the position of target ship θ, dCPA,

tCPA. So the target factors’ discourse domain is:

{ }

tCPAdCPAdu ,,,

The allocation of target factors weight is:

),,,(

tCPAdCPAd

wwwwA

dCPA

tCPA

, and

1=+++

tCPAdCPAd

wwww

Expert recommend:

12.0=

38.0=

dCPA

38.0=

tCPA

Target evaluation matrix is:













tCPA

dCPA

(1)

;10;10;10;10 ≤≤≤≤≤≤≤≤

tCPAdCPAd

rrrr

tCPAdCPAd

rrrr ,,,

are target risk membership.

481

Figure1 . The diagram of evaluation system of ship maneuvering

Distance risk membership function is:

lmm

ddd

dddddu

≤<

≤











−−=

)]/()[(

)(

(2)

distance of the last minute avoidance

distance of adopt avoidance action

DLAKKKd

⋅⋅⋅=

321

RKKKd

⋅⋅⋅=

321

decided by visibility,

decided by water area status,

decided by human factor,

DLA

distance of the last minute action,

is the radius of arena.

)3600(,)]19(cos89.24.4)19cos(7.1

2 

<≤−++−=

θθθ

(3)

Position of target ship membership function:











<≤

−

<≤





360180 ,

)

360

(

1800 ,

)/(1

)(

θθ

(4)

is according to the velocity ratio K of local ship

and target ship

(5)

180













(6)

dCPA risk membership function:











≤<

−

≤

)],

(sin[

)(

dCPAdCPA

dCPA

dCPAu

(7)

milen 1dCPA

)(2

0 t

LL +=

LL ,

are the length of local and tar-

get ship.

tCPA risk membership function:

)(

ttCPA

ttCPAt

ttCPA

tCPAt

tCPAu

≤<

≤











−

(8)

)(

−

(9)

dCPAd

)(

−

(10)

According to the fuzzy comprehensive evaluation

method.

tCPA

dCPA

tCPAdCPAd

wwwwBACR

⋅=⋅= ),,,(

(11)

Collision risk is:

GPS positioning

system

ARPA radar

Integrated mete-

orological meter

AIS

Auxiliary observation

system

Speed meter and

depth meter

Target ship

information

ship

operation

Hydrological

and weather

Screen, sound,

light etc. prompt

Emergency situation

auto decision making

model

Dynamic display

Local ship

information

Navigator’s

operation

482

][

tCPAtCPAdCPAdCPAdd

uwuwuwuwCR +++=

θθ

(12)

4 RESULTS

In a water area, local ship: course 000°, velocity 15

kn, length 75m, the visibility is better(K1=1, K2=1,

K3 =1), adopt DLA=1 n mile. Get the data from

ARPA, target ship: position θ=29.5°, distance d=3 n

mile, relative velocity vs=26.2 kn, dCPA=0.4 n mile,

tCPA= 7 min, length of target ship 110m. calculate

the collision risk of target ship against local ship.

According to the data and associative formula, we

can obtain:

u(dCPA) = 0.8500,

u(tCPA) = 0.3633 ,

u(θ) = 0.6477,

u(d) = 0.1624.

Divide the collision risk into 5 level:

− I—— 1.00～0.91

− II—— 0.90～0.81

− III—— 0.80～0.71

− IV—— 0.70～0.61

− V—— 0.60 ～0.51

According to this division, 0.56 belong to IV lev-

el, middle danger. At this moment, the evaluation

system will display this for navigator. Navigator will

adopt suitable measures according to the information

and self judgment. The evaluation will calculate the

encounter status of two ships in real time. The sys-

tem will send an alarm to navigator for correcting it

when navigator adopts irrational operation. If there

is not any response at the point of last helm, the sys-

tem will adopt automatic collision avoidance strate-

gy.

5 CONCLUSION

Navigation is human’s job, human factor have finali-

ty affect to navigation safety, especially human’s

fault, and it is one of the major reason of shipwreck.

Human’s fault is unforeseen and unconquerable

completely, so we must adopt additional precautions

to improve and make up the affect of human’s fault

to navigation safety. With the development of in-

formation technology, computer is an advantageous

auxiliary facility. Human coordinate with computer

by constructing intelligent evaluation system of ship

maneuvering which makes up human weakness and

also solves the problem that computer is not adapta-

ble to environment. It makes use of the advantage of

human’s adaptability and computer’s calculation ca-

pacity. Therefore, this is a man-machine associative

method, and it is advantageous instrument in naviga-

tion.

ACKNOWLEDGEMENT

This paper is supported by “National Basic Research

Program of China”(No.2008CB417215). The Project

name is “Research on mechanism of shipping safety

and accident-forecasting theory”.

REFERENCES

Davis P. V. et al. 1980. A Computer Simulation of Marine

Traffic Using Domains and Areas, The Journal of Naviga-

tion 33(2): 215-222.

Goodwin E. M. 1975. Statistical Study of Ship Domains, The

Journal of Navigation 28: 329–341.

Guedes Soares C. & Teixeira A.P. 2001. Risk assessment in

maritime transportation, Reliability Engineering and System

Safety, 74: 299-309.

Gaarder S. et al. 1997. Impact of human element in marine risk

management, Advances in safety and reliability pp. 857-

868.

Fuji J. & Tanaka K. 1971. Traffic Capacity, The Journal of

Navigation 24(4): 543–552.

LI Li-na et al. 2006. A Summary of Studies on the Automation

of Ship Collision Avoidance Intelligence, Journal of Jimei

University (Natural Science), 11(2):188-192.

Imazu H. & koyama T. 1984. Determination of times of colli-

sion avoidance, The Journal of Japan Institute of Naviga-

tion 70: 30-37

Kearon J. 1979. Computer programs form collision avoidance

and traffic keeping, Conference on mathematical aspects of

marine traffic. London, pp. 229-242

Thomas Statheros 2008. Gareth Howells and Klaus McDonald-

Maier, Autonomous Ship Collision Avoidance Navigation

Concepts, Technologies and Techniques, The Journal of

Navigation, 61: 129–142.

WU Zhao-lin & ZHENG Zhong-yi 2001. Time collision risk

and its model, Journal of Dalian Maritime University,

27(2): 1-5.