781

1 INTRODUCTION

With transportation modes progressing towards

advanced human-machine teaming and eventually -

autonomy, the topic of handing over control between

agents remains critical. Among the benefits of

autonomous vehicles, one can mention improved

performance in relatively low-complexity

environments, such as highways, railroads, or open sea

navigation [1]. Therein, it is postulated that an

advanced array of environmental sensors and data

analysis algorithms can provide a vehicle with

sufficient situation awareness (SA). However,

unexpected events can still occur there of such a nature

that a vehicle may not necessarily be capable of

handling them on its own and may eventually require

driver’s assistance [2,3]. The vehicle itself can also leave

its Operational Design Domain (ODD) at some point

and by that enter an environment in which it cannot

safely drive itself [4][5]. So far, there are hardly any

systems that can operate completely independently

from their operators - even high-end deep-space

probes do require some intervention from time to time

[6]. It is therefore a matter of time until an autonomous

vehicle requires the driver’s assistance in handling

(un)expected control handover [7]. That raises two

basic questions: (1) what circumstances do push the

vehicle out of its ODD? and (2) how much time is

needed for an operator to safely assume control? Using

expert knowledge, the herein study aims to answer

these questions for a specific type of vehicle, which is

ocean-going merchant vessel.

Time Budget for Merchant Ship Control Takeover -

Preliminary Results

K. Wróbel, K. Formela, P. Kołakowski & M. Gil

Gdynia Maritime University, Gdynia, Poland

ABSTRACT: The increasing automation of shipping requires a proper understanding of the behavior of those in

command of ships to be able to correctly mimic and interpret their actions. This will be especially important

during the widespread introduction of autonomous merchant vessels, whose decision-making algorithms will

need to be correctly prepared to assess situational awareness and will allow for timely control takeover in a

variety of circumstances, including mixed navigation conditions. Therefore, this research aims to investigate the

safety-critical situations when the assistance of a captain is required by the watch officer. There can be various

situations when such a person is called and asked to proceed to Bridge and exercise his/her experience to help

the younger and less skilled colleague in a potentially dangerous situation. In the study, we asked experienced

Masters Mariners about their perception of such a situation. We investigated whether it depends on their sea-

time experience and other factors (e.g. traffic density, weather conditions, fatigue) in a particular situation on

board the vessel. The results of this study may prove valuable in determining the time required for obtaining a

situation awareness during control takeover in different situations. The collected results may also prove useful in

designing and developing navigation simulator exercise scenarios in the context of assessing situational

awareness or providing control-taking guidelines for Maritime Autonomous Surface Ships (MASS).

http://www.transnav.eu

the International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 19

Number 3

September 2025

DOI: 10.12716/1001.19.03.10

782

The issue of determining the time budget [8] for a

takeover procedure is critical for ensuring the safety of

automated vehicles. Ever since a highway hypnosis

hypothesis was formulated [9], it has been widely

accepted that humans are poor supervisors of highly

automated systems [10]. Their minds drift away from

the supervisory task that does not require constant

interaction with a controlled system. With situational

awareness degrading in time, it cannot be assumed that

humans can be brought back to the loop in an instant

[11]. The question therefore arises: if the time required

for a safe takeover is not zero, then what is it? Some

studies suggest values of around or more than 20

seconds [11,12], but they assume a physical presence of

a taking-over actor in the control room. This may not

always be the case, for instance when considering a

Maritime Autonomous Surface Ship (MASS) of such a

design that an autonomous control system requests a

takeover from onboard personnel rather than from a

Remote Operator (RO). In this concept, a reduced crew

is kept on board with no obligation to monitor the ship

constantly but is expected to intervene upon being

prompted. Noteworthy, this setup is similar to that of

contemporary Unattended Machinery Spaces where

duty engineers can rest in their quarters while on

watch.

In our study, we attempted to answer the above

questions indirectly, by asking subjects about their

previous experience in taking conn from a somewhat

autonomous (from their perspective) system - a ship.

2 MATERIALS AND METHODS

For the herein study, we deem a regular merchant ship

as autonomous, as the primary target of our research

were the heads of a deck department, namely Master

Mariners, who are also referred to hereinafter as

seniors. We have asked them about their experiences of

being called to the navigational bridge by a more junior

officer. Such a call can be triggered by a variety of

events as listed in relevant operational procedures,

ranging from safety-critical ones (receiving distress

signals, close-quarters situation developing) to

commercial ones (for instance, new orders received)

[13]. From the seniors’ perspective, the system (a ship)

has been operated autonomously: without interaction

with them, until such interaction was required. Once it

happens, they are required to gain necessary

information pertaining to the reason for being

prompted without prior knowledge of the rationale

behind the call and a degree of urgency and do so as

quickly as possible. They must then assume the control,

and steer the system in a direction required by

circumstances. Needless to say, there is a significant

pressure on them especially in the initial part of the

process when they need to evaluate not only how time-

critical it is to decide on the further course of action but

also to make sure that the decisions made are actually

correct ones. There can also be a case when juniors call

them too late to take any decisive action, just as may

happen when self-driving cars alert the driver in

insufficient advance.

To collect seniors’ experiences about the master to

the bridge situations, we set up an online

questionnaire. To ensure the relevance and accuracy of

the survey, it was developed in collaboration with

marine officers, including experienced Master

Mariners. These experts were consulted to identify key

situations that typically require a handover of control

on board ocean-going merchant ships. Their input

helped in designing questions that accurately reflected

real-world scenarios, such as emergency calls to the

bridge, technical failures, adverse weather conditions,

and collision avoidance.

An invite has been sent to various shipping

companies, seafarers’ associations, etc. to promote it.

Additionally, a request to invite more respondents was

included which makes a response rate impossible to

calculate. Eventually, 47 individuals (all were males

and held the Master Mariners licenses) filled out the

questionnaire. Their demographic breakdown is

presented in Figure 1.

Figure 1. Demographic data of the respondents

3 RESULTS

3.1 Overview

Of 47 Master Mariners elicited, 38 declared having

been called to the bridge in an emergency on an

average of about 11 times during their career span

(ranging from 1 to 30). Then, the respondents were

requested to list the most frequent reasons for which

they have been called to the bridge. The breakdown of

the received answers is depicted in Figure 2.

Figure 2. The most frequent reasons for which senior officers

have been called to the navigational bridge

783

In the next step, participants were asked to estimate

the time needed for them to reach the bridge and gain

situation awareness to the point, in which they would

be ready to make necessary decisions. On average, they

declared some 2.9 minutes in emergencies and 6.1

minutes in routine circumstances. A histogram of their

answers is given in Figure 3. Noteworthy, the time

declared by subjects as required to reach the

navigational bridge and gain the SA showed a near-

zero (-0.007) Pearson correlation coefficient with their

experience: both in rank and in total. This indicates that

with respect to time budget management, experienced

captains do not seem any better than inexperienced

ones, at least declaratively.

Figure 3. Time needed to reach the navigational bridge and

gain situation awareness

In the further part of the study, the respondents

were asked to recall their most recent emergency call to

bridge and the most stressful one. These questions

were not required in the survey, and thus have not

been answered by some of the respondents, so the

sample sizes can vary.

3.2 Most recent experiences

Over 90% of the most recent calls to Master have been

related to either collision avoidance, technical failures,

or weather conditions. The respondents declared that

it took them some 3.2 minutes (on average) to build the

situation awareness and take the conn from the

moment they were called. Their time budget amounted

to an average of 10.3 minutes. They also stated that the

handover by an OOW was effective (32 responses) or

partly effective (5 responses). However, in 9 cases they

mentioned that the bridge alarm system did not help

gain their situation awareness. This could indicate that,

for some reason, the subjects perceived human-to-

human information exchange as more effective than

the machine-to-human one.

However, the most recent experience does not give

a relevant picture of situations in which it would be

critical to execute a smooth and efficient handover.

These would rather consist of situations that have been

found most stressful by the subjects.

3.3 Most stressful experiences

From the safety perspective, it can be raised that the

more stressful the situation is, the more disastrous can

be its potential consequences. Therefore, a level of

stress felt by operators can serve as an indicator of how

dangerous the circumstances might have become if not

for mitigating actions taken. With this in mind, it is

necessary to take a closer look at situations in which the

efficient and safe hand-over was critical for the safety

of the vessel involved. As can be seen in Figure 4, the

breakdown of these most stressful situations is to some

extent similar to the overall breakdown of situations, in

which master’s assistance was required (see Figure 2).

However, a few interesting points can be noted:

− Traffic and navigation-related calls were in general

less stressful (challenging) even though they were

quite frequent.

− Situations related to collision avoidance were the

most frequently labeled as most stressful while their

frequency was similar to that of Technical and

Traffic/navigation. It is of note that captains

(respondents in this study) are by default the most

experienced nautical officers on board their vessels

and they were still the most stressed in collision

avoidance situations, ones that they have likely

personally resolved countless times in their careers.

This, in turn, may indicate that they have been

called by OOWs too late into ships encounters. It

can also be noted that situations related to collision

avoidance are caused by direct actions of some

humans - crewmembers of both Own Ship and

Target Ship(s). These are unlike other factors

leading to frequently stressful calls to Master, such

as Technical breakdowns (which can be random or

can result from factors with which nautical officers

are usually less familiar) and Weather conditions,

which are natural phenomena outside of human

control.

− Respondents indicated that Fire on board was

among the most stressful situations in which they

were called to the bridge. This answer was not

predefined in the survey and even the respondents

who provided them here did not mention Fire in

previous, open-answer questions. By this, the

respondents did not associate fire with situations

they would be normally expected to handle and

focused on those related to their more routine

duties (navigation, collision avoidance,

communication) when asked a routine question. It

was not until they were asked specifically about the

most stressful (challenging) event they have found

themselves in that they mentioned a rare but

extremely dangerous event.

The respondents declared that it took them some 3.1

minutes to arrive at the bridge and build the situation

awareness from the moment they were called to assist

the OOW, which is similar to the times declared

previously. However, they also declared a time budget

of 8.1 minutes - some 21% less than in the most recent

situations (10.3 minutes). The latter are not necessarily

highly challenging (some of them are, but others are

not). The fact that it took Masters a similar time to

respond to the call in the most stressful and ‘average’

situations indicates that they react to the call by an

OOW in a similar way. Presumably, it would take too

much time for a caller to explain all circumstances to

the captain and specify the level of urgency. Masters

would rather respond to the call as quickly as possible

and determine the urgency only upon arriving at the

conning station. Time budget can be named as a proxy

of such urgency - the less time is available to avert a

hazard, the more urgent is master’s intervention. 10.3

minutes on average, 8.1 minutes when the situation

drifts into hazard.

784

Figure 4. Most stressful reasons for calling Captains

according to the respondents

The control takeover process was evaluated more

skeptically than in an average situation. Handover by

an OOW was partly effective in 9 of 37 cases and

ineffective in only one case. Messages and alarms

produced by bridge automation were partly effective

in eight cases and ineffective in 11 of 37 cases (30%).

This indicates that particularly in stressful and

demanding situations, messages automatically

generated by devices installed on the navigational

bridge are of limited help. As the purpose of these

messages is to improve operators’ situation awareness

by attracting their attention to the potentially

dangerous factors, the fact that they are not as useful as

should be may indicate a need to re-think their design.

This is particularly important for automation-to-

human handovers in non-routine situations. Herein,

the person taking the sole conn of the MASS would

have to rely only on information fed by the automation,

at least in the initial phase of take-over. They would not

have the privilege of human-to-human hand-over,

which allows for an exchange of non-standard

information. Note that OOW handing over to master

was evaluated as effective in 27 of 37 most stressful

situations while the bridge alarming system was only

effective in 18 of them.

4 DISCUSSION

In the below subsections, the outcomes of the

performed study are discussed along with its

limitations.

4.1 Findings

The performed study helped investigate situations in

which experienced Master Mariners have been called

to assist their less experienced colleagues in potentially

hazardous or otherwise non-standard situations. It also

made it possible to assess the time required to gain

situational awareness as well as to take control of the

ship. The entire master to the bridge call along with the

study results discussed in the subsequent paragraphs

are graphically summarized in Figure 5.

Firstly, a breakdown of types of such situations and

their levels of urgency have been identified. Most of the

situations in which masters are called to the bridge are

related to technical malfunctions, deteriorating

weather, and traffic conditions including a direct threat

of collision. Among the less frequent answers that may

be relevant for MASS are: receiving a distress signal

and dragging anchor. From a perspective of MASS

design, it must be noted that the list is not exhaustive

and that the assistance of a human operator (not

necessarily a certified Master Mariner) may be

necessary in even less standard and unexpected

circumstances. To this end, the case of Fire as outlined

above serves as a good example. Herein, respondents

did not list Fire as a reason for being called to bridge,

until asked about the most stressful situation in which

it so happened to them. This indicates a need for a

careful design of MASS control systems, taking into

considerations the knowledge of properly and

thoroughly elicited, experienced seafarers. After all,

how many non-standard, infrequent yet safety- or

security-critical situations can be out there at high seas?

Not in the handbooks, procedures or accident reports,

but living memories of professionals.

Secondly, the declared times in which Masters gain

situation awareness and react to the circumstances met

have been studied. It has been found that masters reach

the bridge and obtain a situation awareness in some 3

or so minutes on average. This does not appear to

depend on the urgency of the situation in which their

assistance was found necessary or required by

operational procedures.

It is of note that maximum declared times, being the

most critical from the safety perspective, were in some

cases (see Figure 3) several times larger than obtained

mean values.

Figure 5. Graphical summary of the ‘master to the bridge‘ procedure during the navigation process and its implications for

situational awareness and control takeover, inspired by [8]

785

The outlying ones reached even up to 30 minutes

for routine and 10 minutes for emergency master to the

bridge call. Interestingly, quite a few of the captains (9

of 47, i.e. almost 20% of the sample) reported they were

able to reach the bridge and gain SA in 5 minutes in

case of emergency. A total of 8 of them declared that

their last offshore assignment had been done on board

of container vessel, which is a rather fast ship type

compared to other merchant ones. Thus,

conservatively assuming 18 knots as a representative

speed, it means that between calling the master and his

response, the ship advances some 1.5 NM and can

eventually get significantly closer to the potential

threat (e.g. during a close-quarters situation).

Consideration of these times of control takeover on

manned ships seems to be of utmost importance to

ensure proper response of MASS during mixed traffic

conditions when an autonomous ship will meet and try

to mimic the behavior of a manned one or at least

correctly interpret the development of ship encounter.

However, it must be noted that the declared time

required to reach the conning station and gain the

situation awareness does not account for the fact that

masters need to physically move from wherever they

are to the bridge. This may be different from the MASS

Remote Operations Centre setup where operators

would likely be seated in a direct vicinity of their

respective control consoles.

4.2 Limitations

Among the limitations of the current study, potentially

affecting the credibility of its results, the following

issues can be listed:

− the questionnaire was filled out by 47 Master

Mariners. Only 37 of them responded to all the

questions. Given the fact that there are some 50,000+

ships registered globally [14] and likely twice as

many certified Master Mariners, our sample can

hardly be called representative. However, the

answers provided by even such a small sample do

indicate some factors that may be found relevant for

the development of prospective MASS and can be

used to bring attention to these factors;

− the performed study was based on respondents’

declarations rather than solid experimental data.

Respondents were requested to retrieve certain

facts from memory, which can be misleading.

5 CONCLUSIONS

This study sheds light on the critical aspects of control

takeover in maritime settings, particularly for ocean-

going merchant ships. The findings emphasize the

importance of timely and effective human intervention

when unexpected situations arise. Our findings

highlight the significant variability in response times

and the nature of situations that necessitate the

intervention of experienced personnel, such as Master

Mariners, when control is handed over from junior

officers.

It was found that the most common triggers for

emergency calls to the bridge were collision threats

(44%), technical malfunctions (25%), heavy traffic

(21%), and adverse weather conditions (14%). The

survey results indicate that captains generally require

about 3 minutes to respond to an emergency call and

achieve sufficient situational awareness to take control

of the vessel. This response time appeared relatively

consistent regardless of the urgency or stress

associated with the event, suggesting a standardized

approach by masters to emergencies.

However, the study also reveals a significant gap in

the effectiveness of automated alert systems compared

to human-to-human communication during

handovers. In high-stress situations, automated

systems were often found to be less effective in

conveying critical information, suggesting that current

designs may not adequately support decision-making

under pressure. This limitation points to a need for

enhancing the design of these systems, especially as the

maritime industry moves towards greater automation

and the development of Maritime Autonomous

Surface Ships (MASS).

The insights gained from this research highlight the

ongoing necessity for human oversight in automated

systems and the need for further refinement of control

handover processes. Future advancements should

focus on improving the reliability and intuitiveness of

automated alerts and ensuring that human operators,

whether on board or remote, can seamlessly take over

control when needed.

ACKNOWLEDGMENT

The herein study has been supported by Gdynia Maritime

University internal grants #WN/2025/PZ/12 and

#WN/2025/PZ/01.

The authors express their sincere gratitude to the

anonymous participants of the study.

REFERENCES

[1] N. Merat, A.H. Jamson, F.C.H. Lai, M. Daly, O.M.J.

Carsten, Transition to manual: Driver behaviour when

resuming control from a highly automated vehicle,

Transp. Res. Part F Traffic Psychol. Behav. 27 (2014) 274–

282. https://doi.org/10.1016/j.trf.2014.09.005.

[2] K. Bogusławski, J. Nasur, J. Li, M. Gil, K. Wróbel, F.

Goerlandt, A Cross-Domain Scientometric Analysis of

Situational Awareness of Autonomous Vehicles With

Focus on the Maritime Domain, IEEE Access. 10 (2022)

50047–50061.

https://doi.org/10.1109/ACCESS.2022.3174097.

[3] C. Correa-Jullian, M. Ramos, A. Mosleh, J. Ma, Task

Allocation And Control Transitions In Autonomous

Driving System Operations, in: ESREL 2024 Monogr. B.

Ser., Kraków, 2024: pp. 77–86.

[4] NHTSA, A Framework for Automated Driving System

Testable Cases and Scenarios, Washington, DC, 2018.

[5] P. Weissensteiner, G. Stettinger, S. Khastgir, D. Watzenig,

Operational Design Domain-Driven Coverage for the

Safety Argumentation of Automated Vehicles, IEEE

Access. 11 (2023) 12263–12284.

https://doi.org/10.1109/ACCESS.2023.3242127.

[6] C. Cofield, NASA’s Voyager Will Do More Science With

New Power Strategy, NASA. (2024).

https://www.jpl.nasa.gov/news/nasas-voyager-will-do-

more-science-with-new-power-strategy (accessed

September 3, 2024).

[7] T. Gruden, J. Sodnik, G. Jakus, Assisted Partial Take-Over

in Conditionally Automated Driving: A User Study, IEEE

786

Access. 11 (2023) 107940–107957.

https://doi.org/10.1109/ACCESS.2023.3320286.

[8] B. Zhang, J. de Winter, S. Varotto, R. Happee, M. Martens,

Determinants of take-over time from automated driving:

A meta-analysis of 129 studies, Transp. Res. Part F Traffic

Psychol. Behav. 64 (2019) 285–307.

https://doi.org/10.1016/j.trf.2019.04.020.

[9] G.W. Williams, Highway Hypnosis: An Hypothesis, Int. J.

Clin. Exp. Hypn. 11 (1963) 143–151.

https://doi.org/10.1080/00207146308409239.

[10] B. Strauch, Ironies of Automation: Still Unresolved After

All These Years, IEEE Trans. Human-Machine Syst. (2017)

1–15. https://doi.org/10.1109/THMS.2017.2732506.

[11] T. Cheng, E.A. Veitch, I. Bouwer, M.A. Ramos, A.

Mosleh, O. Andreas, B. Wu, Analysis of human errors in

human-autonomy collaboration in autonomous ships

operations through shore control experimental data,

Reliab. Eng. Syst. Saf. 246 (2024) 110080.

https://doi.org/10.1016/j.ress.2024.110080.

[12] O. Shyshova, P. Gadhavi, M. Tenzer, F. Tanshi, D.

Söffker, Takeover time: Requirements for highly

automated inland vessels - First experimental-based

results, in: 2024 IEEE 4th Int. Conf. Human-Machine Syst.,

IEEE, 2024: pp. 1–7.

https://doi.org/10.1109/ICHMS59971.2024.10555860.

[13] International Chamber of Shipping, Bridge Procedures

Guide, Fifth, Marisec Publications, 2016.

[14] UNCTAD, Review of maritime transport 2022, Geneva,

2022. https://unctad.org/system/files/official-

document/rmt2022_en.pdf.