International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 2

Number 1

March 2008

Fairway Navigation – Observing Safety-Related

Performance in a Bridge Simulator

R. Nilsson & M. Lützhoft

Chalmers University of Technology, Göteborg, Sweden

T. Garling

Göteborg University, Göteborg, Sweden

ABSTRACT: This paper proposes an approach of measuring navigation performance using a full mission

bridge simulator. The motivation for this research is the updates in equipment and that the desire of using new

instruments and technology not always is accompanied by analyses of the impact of the changes. The task of

navigating in a fairway is proposed to be assessed through various methods to answer questions related to

performance and the experience of using bridge equipment. The overall aim is to reach a higher degree of

understanding and knowledge through the testing of different instrumentation setups.

1 INTRODUCTION

The equipping of a ship bridge has during the last

decades changed substantially. One reason for this is

the development of instruments such as for instance

the Global Positioning System (GPS), electronic

charts systems and the Automatic Identification

System (AIS). This development in combination

with an eagerness to adopt new technology and

incorporate it into ship bridge systems contributes to

the vast variety of instruments and systems on a ship

bridge. Governments and companies are naturally

from a safety and efficiency perspective interested in

using new technology. An investigation of piloting

in Sweden was started by the Swedish Ministry of

Enterprise, Energy and Communications in December

2006. The overall aim of this investigation is to

review the possibilities of using new technology.

One issue to be investigated is “shore-based

pilotage” and whether piloting with the help of new

technology could be more efficient through support

from a shore-based central.

As shown in an earlier study of the accident

involving the ship “Royal Majesty” (Lützhöft, 2002)

new technology is not necessary equivalent to higher

safety. It depends on how well the operators know

the technology and its constraints. Some radar

training instructors identified new behaviors during

simulator training which violated existing rules for

navigation (Lee et al., 1993). These violations were

believed to be triggered by the experienced

reliability of the new equipment.

The seafarer is remarkable at adjusting to new

circumstances (Lützhöft, 2004). In this context

Lützhöft discussed the positioning of equipment on a

ship bridge. On a traditional bridge, where all the

equipment is placed in one row, and where new

equipment has been added after years of sailing, it is

not unusual to find complemented equipment (like

for instance an Electronic Chart System) placed

where there was room for it, usually far at one end.

This could be compared to a modern cockpit bridge

design where the Electronic Chart System has

a central position in the bridge layout.

A study of pilots in Finnish coastal waters

(Norros, 2005) showed that personal piloting style

affects the way that the piloting task is solved more

than the available technology at the ship bridge. This

indicates that a key factor which has to be taken into

account is the way the equipment is used.

In order to understand how new technology

affects navigational performance in fairways, it is

essential to gain more knowledge of how the ship is

navigated.

2 THE PROBLEM

Taking a ship to and from a berth always involves

some safety risk. According to Boisson maritime

safety is both the material state resulting from the

absence of exposure to danger, and the organization

of factors intended to create or perpetuate such a

situation” (Boisson, 1999, p. 31).

The Swedish Ministry of Enterprise, Energy and

Communications wishes to know how new

technology can be used to facilitate piloting and

make it more efficient. The ministry also wants to

know what the prerequisites for developing “shore-

based pilotage” are.

In order to know more about how new technology

can be used in the future, our aim is to learn more

about how work is performed at present and what the

prerequisites for carrying out the task are. One way

to gain more knowledge is to compare two sets of

equipments that are presently available on a ship

bridge which reflect differences due to the changes

in technology that have already taken place.

Questions raised include: How is the work

experienced on the ship bridges with existing

technology? Are there any differences in experience

and performance due to differences in equipment

standard? More specifically, we want to:

1 Measure the experience of workload related to

two different sets of equipment.

2 Compare the experienced feelings related to the

work situation.

3 Compare the performance in the task of

navigation in fairways related to sea safety.

We are also interested in analyzing the

performance on the bridge to find if there are any

salient navigation strategies that are manifested in

one or both of the tested work environments.

3 METHOD

We suggest that learning more about the work on a

ship bridge can be obtained through studies of work

in a full mission bridge simulator. “Full mission”

means that the environment where the navigation

task is simulated is authentic in comparison to

equipment that could be found on an operating ship.

3.1 The value of simulator studies

As long as the tasks are realistic and the performance

can be analyzed so that it is possible to separate its

determinants, simulator studies are valuable. Funke

(1988), who used simulations to study complex

problem solving, stressed that “it should be analyzed

how participation in simulation affects problem

solving in ‘real’ life problem situations” (p. 297).

For instance, many cruise companies stress that the

use of navigation simulators in training is a way to

enhance performance. Navigating a ship is basically

a dynamic decision making task. According to

Brehmer (1999, p 10) such tasks have three

important characteristics:”

− They require a series of interdependent decisions;

− The state of the task changes, both autonomously

and as a consequence of the decision makers

actions;

− The decisions have to be made in real time.”

These characteristics of ship navigation can be re-

created and evaluated in a full-mission simulator by

having participants solving tasks that are realistic,

representative and carefully designed.

3.2 Tasks

The task of navigation can differ depending on the

ship, e.g., factors like size and propulsion capacity,

and the area, e.g., the water to be navigated.

Normally out in the open sea there is no need for

piloting. When modeling the pilot task, Norros

(2004) divided piloting into two different types of

piloting called sea piloting and harbor piloting. Sea

piloting refers to the navigation through the

archipelago and/or fairways, and harbour piloting

refers to the “maneuvering of the ship in the harbour

area” (p. 186). Based on the special interest in

piloting from the Swedish Government, the focus of

this research will be navigating in fairways,

comparable to the one Norros refers to as sea

piloting. In the simulator participants will be asked

to solve navigation tasks in confined waters.

3.3 Understanding and creating the task

In order to create an understanding of the task to be

studied and lay a foundation for the creation of the

scenarios, a Hierarchical Task Analysis (HTA) was

conducted. The HTA was based on interviews with

four experts, fully authorized marine pilots, lasting

approximately 2 hours each. Of the four experts

interviewed three are still working as pilots. The

interviews were conducted at two different locations

at various times of the day. Pilots were chosen for

this part of the interview because they naturally and

frequently change ships in their work. This gives

them experience of various types of ships which was

considered as favorable. When the task of navigation

in fairways was decomposed to a satisfying level, the

work of finding problems related to the task began.

With the HTA functioning as a foundation,

instructors at a maritime university and active

captains were interviewed. This time the focus was

on the perspective of the captain in piloting

situations. Discrepancies in communications

between pilot and captain have been found in some

studies.

In total, four captains in active duty and two

instructors were interviewed. Of these, two captains

and two instructors were interviewed in a group and

two captains individually. The interviews lasted

approximately two hours.

The results from the HTA and the problem

interviews served as a foundation for creating

representative scenarios for the tests in the simulator.

3.4 The test setting and participants

Two configurations of bridge types were chosen.

One setting called “traditional bridge” consists of

less advanced technology. The second bridge is an

advanced bridge with an Integrated Navigation

System (INS). Both bridge types will follow

regulations regarding what equipment that is

required. In table 1 the most evident differences are

presented.

Table 1. Major differences between the bridge types

Traditional bridge Integrated Navigation Bridge

Paper Chart

Electronic Chart System, although

paper chart available

Automatic Information

System through a

Minimum Keyboard

Display

Automatic Information System

integrated with the Electronic

Chart System and/or Radar

Basic function on

autopilot, no function

like “curved headline”

Advanced autopilot with the

function “curved headline”

No conning display,

burt information

available elsewhere

Conning display

Requirements to plot

position in paper chart

No requirement to plot position in

paper chart

Possibilities to overlay

information systems like

Electronic Chars System and

Radar

We plan to test 28 bridge crews consisting of two

members each. Each bridge crew will conduct test

scenarios on both ship-bridges. This is exemplified

in table 2 where two scenario trials for two bridge

crews, α and β, is exemplified.

Table 2. Participation over time for two different bridge teams

Bridge

Team

Participation

order

Bridge Type Scenario

α 1

Traditional

Bridge

INS Bridge

β 2

Traditional

Bridge

The route will be preplanned and each team will

have time to familiarize themselves with the route.

The task is to navigate the ship according to the

route plan as fast as possible with maintained safety.

Two pilot studies have been conducted in which

both settings and scenarios were tested. Two active

captains and three cadets, soon to be 2nd officers,

tested the scenarios and bridge settings. The

response was that the scenarios were realistic and

that they were equal in difficulty. Although perhaps

not all the events could be expected to be

encountered on one passage, the response to the pilot

studies were positive.

Participants for the study are being recruited

among active captains and cadets in their final year,

soon to be 2nd officers.

4 WHAT TO MEASURE, WHY AND HOW?

4.1 Risk and Sea Safety

Recommendations from the bridge procedures guide

and STCW regulations will play a central part for the

assessment of work performance on the ship bridge.

This assessment will be done by experts.

4.2 Performance

During the voyage the route will be registered so that

it will be possible to compare the planned route with

the performed one. A debriefing session will be held

after each trial, during which reasoning about

deviations from the planned route regarding for

instance speed and heading will be discussed.

Judging performance related to sea safety is a

complex matter. In some studies Cross Track Error

(XTE) has been used as an indicator of performance.

We argue that XTE in our case is less accurate from

the perspective of sea safety. To follow a planned

route in detail is not necessarily safer than to deviate

for safety reasons, therefore we consider it

unrealistic to ask the captains to focus on staying on

the track as the main task. It could lead to

participants not navigating as they would usually do,

in order to try to stay on track. Instead we use

experts to rate each bridge teams’ performance

according to existing rules and procedures. A safety

margin will however be assessed at a number of

given points.

We will register average speed since it is valuable

to use the speed to assess sea safety.

It is interesting to measure Closest Point of

Approach (CPA) in some cases. CPA is also a

measure that has to be dealt with carefully to not

overestimate potential danger. A ship can pass

another ship by the stern with a relatively small CPA

and still have acted in a safe manner. The same CPA

could be more dangerous if the ship was passed on

the bow.

4.3 Workload through heart rate

Heart Rate Variability has been used to measure

mental workload as “variability is generally found to

decrease as the load increases” (Wickens, 2000,

p.465). In aviation research, tests have been

conducted to compare reactions in the real world

with reactions in simulators(Magnusson, 2002).

The results indicate that the psychophysiological

reaction patterns for the two settings are very

similar. We expect to attain a good measure of

mental workload at the ship bridge through

measurements of HRV.

4.4 Subjective measures

We will use three subjective measures to collect

information regarding participants’ experience of the

work setting.

4.4.1 Subjective Task Load

We will use the NASA Task Load Index (TLX)

as a self-reporting method for assessing workload,

both physical and mental. When using TLX,

workload is defined as the “cost incurred by human

operators to achieve a specific level of performance”

(Gawron, 2000, p.130). We are interested in

measuring workload that is experienced during the

task. This can be used as a control of the scenarios as

it will provide information regarding the workload

experienced in the bridge settings. The participants

will be asked to rate the task on six different

dimensions after having performed the tasks in the

simulator.

4.4.2 Affective responses

The Swedish Core Affect Scale (SCAS) is a

method which is used as a self-report measure of

core affects. It measures core affects which are

“cognitively accessible elements of a current mood”

(Västfjäll, Friman, Gärling & Kleiner, 2002, p.19).

This means that measurements from the same

individual can differ depending on the current mood.

In our study participants will assess their mood on

two dimensions immediately before and after the

task solving in the simulator. The two dimensions

are valence and activation. “The valence dimension

is interpreted as reflecting the degree of affect that

provides information about the current well-being

[…] activation, refers to subjective experience of

energy or mobilization” (Västfjäll et al., 2002, p.20).

An earlier study shows that the adjective ratings are

“reliable measures of the independent valence and

activation dimensions proposed” (Västfjäll et al.,

2002, p.19). It has also been shown that SCAS has a

positive correlation with Heart Rate Variability. In

this study we will analyze self reports of core affects

to see if there are any differences in the experience

of using the two bridge types.

4.4.3 Experience of control

From a Joint Cognitive System point of view

(Hollnagel, 2005) both the bridge crew and the

bridge equipment are constituents of the system that

we will study.

The participants will answer a question about

their perceived degree of control at regular intervals

along the scenarios. When answering the question

they will describe their experience of the situation as

a whole and thus the experience of working in the

system. By doing this it will be possible to compare

the experience of working at the two bridge systems

as a whole.

5 EXPECTED FINDINGS

5.1 Comparing new and old technology

We expect the results to provide information on the

impact of changing bridge equipment, and how the

workload on the ship bridge is experienced. We will

also learn more about how information is handled

and what kind of information that is represented in a

good way and gain clues to which information

representation can be enhanced.

5.2 Data

We expect the following data to be accessible:

1 Video recordings. The work with the electronic

equipment like radar and ENC (Electronic

Navigation Chart) will be recorded.

2 Expert evaluations of each scenario trial

3 Independent performance measures such as:

− Time (average speed)

− Measures of relevant safety margin

− Measures of relevant CPA

4 Heart Rate (HRV)

5 Experienced work load (NASA TLX)

6 Experienced core affect (SCAS)

7 Experienced control during the scenario trials

These data will serve as a foundation for

comparing the two bridge settings. We intend to

identify and analyse any salient work strategies.

These will be compared to recommendations from a

risk perspective.

6 LIMITATIONS

The task solving in navigating a ship is complex and

dynamic. This study will be based on observations of

performance in a simulator and although we have

argued for the realism in the scenarios, there will

always be doubts regarding the possibility of

generalising our results. We will make the scenarios

as realistic as possible but of course some constraints

that appear unnatural may still have an effect. One

example is the time. The scenarios will last for

approximately one hour. In reality the legs of sailing

are often longer. At the same time the number of

critical events is much more frequent than they

would be naturally.

The models used will have some constraints, as

we rely on the functioning of the technology used,

and factors like for instance how well the ship model

is programmed and interacts with the imaginary sea

can have an effect. The feeling of “this is not for real

anyway” may be present to some degree. Related to

the dynamics of the task one could always question

the possibility of generalising results. However, in a

simulator we are able to describe the prerequisite for

the scenarios in detail, and play back the recordings,

and thus learn something from these situations.

We have a limited opportunity for choosing

participants. In this study active captains and

advanced cadets have been invited to participate.

This may lead to a positive selection. Thus, our

sample may overrepresent participants with a special

interest for this kind of work or for knowledge of the

simulator itself.

7 CONCLUSIONS

We have primarily set out to learn more of the

solving of the task of navigation in fairway

performed at a ship bridge. Our measurements will

cover both user perspective and performance related

to the joint cognitive system. With this diversity of

measurements we believe we have a good

opportunity to increase the knowledge base

regarding possibilities and constraints of new

technology.

ACKNOWLEDGEMENTS

We gratefully acknowledge the financial support

from VINNOVA (Swedish Governmental Agency

foe Innovation Systems). We would further like to

thank the Swedish Maritime Administration for their

support during the work in this project.

REFERENCES

Boisson P., 1999. SAFETY AT SEA Policies, Regulations &

International law. Paris: Edition Bureau Veritas.

Brehmer B., 1999. Reasonable Decision Making in Complex

Environment. In Juslin, P. & Montgomery, H. 1999,

Judgment and Decision Making: Neo-Brunswikian and

Process-Tracing Approaches. (pp. 9-21)London: Lawrence

Erlbaum associates.

Funke J., 1988. Using simulation to study complex problem

solving, Simulation and games, vol 19 No3, September

1988: 277-303.

Gawron VJ., 2000. Human performance measures handbook.

London: Lawrence Erlbaum Associates.

Kommittédirektiv Dir.2006:116.

http://www.regeringen.se/sb/d/8095/a/74338;jsessionid=aWTR

XOif4qz_

Lee JD., & Sanquist TF., 1993. A systematic Evaluation of

Technological Innovation: A case study Of Ship navigation.

Battelle Seattle research Center. Seattle: WA 98105.

Lützhöft M., 2002. On your watch: Automation on the bridge.

Journal of Navigation, 55(1): 83-96.

Lützhöft M.. 2004. The Technology is great when it works.

Dissertation N0.907. Linköping: Linköping Studies in

Science and Technology. University of Linköping.

Magnusson S., 2002. Similarities and Differences in

Psychophysiological Reactions Between Simulated and

Real Air-to Ground Missions. The international journal of

aviation psychology 12(1): 49-61.

Transportation Safety Board of Canada (1995), A safety study

of the operational relationship between ship masters/ watch

keeping officers and marine pilots, Report nr SM9501

Available at: http://www.tsb.gc.ca/en/reports/marine/ studies/

ms9501/ms9501.asp.

Norros L., 2004. Acting under uncertainty, Espoo: VTT

publications.

Wickens CD. & Hollands JG., 2000. Engineering Psychology

and human performance. 3:rd edition, New Jersy: Prentice-

Hall Inc.

Västfjäll D., Friman M., Gärling T., Kleiner M., 2002.The

measurement of core affect: A Swedish self-report measure

derived from the affect circumplex. Scandinavian Journal of

Psychology 43:19-31.