575
1 INTRODUCTION
Artificial intelligence is gradually infiltrating human
life and challenging traditional human paradigms in a
variety of industries, including transportation [33].
Vessel navigational aids have been upgraded, such as
vessel positioning, which has changed from the old
sextant astronomical positioning to GPS, DGPS, and
radar object positioning [27]. However, the sextant is
still on board the vessel in case of emergency, in case
the electronic positioning equipment unexpectedly
malfunctions while the vessel is traveling across
oceans. Although traditional vessel manoeuvring is
controlled by humans, it is gradually undergoing a
transformation. MASS is a prominent concept that
could reshape the current maritime market [43].
The International Maritime Organization (IMO)
started defining MASS in 2017 in response to the fast
growth of MASS and divided it into four levels of
autonomy: crewed vessel with automated processes
and decision support (Degree One); remotely
controlled vessel with seafarers on board (Degree
Two); remotely controlled vessel without seafarers on
board (Degree Three); and fully autonomous vessel
(Degree Four) [16]. But before MASS achieves
complete autonomy, there are still major obstacles to
overcome. One of the major challenges is the
development of a safe collision avoidance system that
is able to comply with traffic regulations while also
safely navigating the sea with other MASS and
traditional vessels [1]. The Convention on the
International Regulations for Preventing Collisions at
Sea, 1972 (COLREGs) refers to the set of navigational
regulations that ensure safety and order at sea, and
Challenges and Solutions to COLREGs on MASS Ethical
Issues in Vessel Manoeuvring Practices
L
. He
Shanghai Maritime University, Shanghai, China
ABSTRACT: Though much attention has been paid to the navigational safety of MASS, very few have examined
the major challenges that MASS presents to COLREGs from the standpoint of practical vessel manoeuvring. The
introduction of MASS may alter the traditional paradigm of vessels depending on seamen for vessel
manoeuvring. COLREGs safeguard the orderly statutory nature of vessel navigation. This paper discussed the
proper understanding of COLREGs and the ethical MASS issues that can arise from vessel manoeuvring
practices in restricted and heavy waters in coastal ports, challenging “look-out,” “good seamanship,” and
navigational safety in COLREGs. The paper also discusses the feasibility of modifying COLREGs and how to
solve these problems without modifying COLREGs. The paper suggests expanding the pilotage distances and
applying the legal system of maritime traffic management in an integrated manner so as to ensure the safety of
MASS navigation in the restricted waters of coastal ports and in heavy waters from the perspective of pilotage
and the legal system. This paper suggests th
at it would be prudent for the IMO not to apply COLREGs directly
to MASS until the navigational safety of MASS has been fully considered.
http://www.transnav.eu
the
International Journal
on
Marine Navigation
and Safety of Sea Transportation
Volume 18
N
umber 3
September 2024
DOI: 10.12716/1001.18.03.1
1
576
the IMO is responsible for formulating modifications
to these regulations. MASS navigation has become
more and more common in various countries, and
complex legal and safety issues have arisen as to
whether or not the rules of the COLREGs apply to
MASS. This is particularly true for remotely operated
vessels without crew (Degree 3) and fully
autonomous vessels (Degree 4). This paper will
therefore focus on the application of the
understanding of COLREGs in relation to Degree 3
and Degree 4 MASS and the challenges of
navigational safety.
The issue is a contentious matter to determine
whether the COLREGs' “good seamanship” and
“look-out” rules directly apply to MASS. Most
academics agree that MASS can meet COLREG's
“look-out” requirements through technological means
[9]. Considering that the COLREGs apply to “vessels”
rather than “masters or crews,” it appears that “good
seamanship” requirements can be met [20]. Some
academics are concerned that technological
improvements have the potential to replace some of
the capabilities of humans, particularly in situational
awareness. Nonetheless, concerns persist regarding
sensors' capacity to outperform the human eye and
ear in terms of perception in order to offer a thorough
evaluation of situational and collision risk [25].
According to recent research, it has been found that
many academics are interpreting a term or a separate
clause in the COLREGs and have not interpreted the
rules from the technical point of view of vessel
manoeuvring practice, nor have they considered that
the MASS involves the issue of intelligent ethics and
the challenge of intelligent ethics to the COLREGs.
With these considerations in mind, the main
objective of this paper is to examine the difficulties
that arise when examining the application of
COLREGs from the technical standpoint of
manoeuvring practices for MASS vessels.
Additionally, this paper offers both operational and
legal solutions to address these issues. The format of
the paper is as follows: In Section 2, the main
challenges posed by vessel manoeuvring practices
with regard to navigational safety and intelligent
ethical issues are analysed, along with the
understanding of ship manoeuvring practices with
respect to the “look-out” rules and “good
seamanship” of the COLREGs. Section 3 explores
various approaches and presents potential operational
options in the future. The discussion is concluded,
and recommendations are given in Section 4.
2 EXPLORING COLREGS AND ETHICAL
CONCERNS WITH MASS
According to statistics, human factors are involved in
80.7% of marine incidents and collisions [12].
Nonetheless, human error is likely to occur in MASS
at degrees 2 and 3, when the impact of the human
component will grow [19]. Meanwhile, remote
operators working ashore, who are unable to directly
observe their surroundings, have less confidence in
the physical constraints and limited visibility that can
diminish the remote operator's ability to make correct
judgments [14]. And MASS has the greatest risk of
navigation and chance of accidents caused by
machine breakdown when sailing at sea [7]. On
Degree 3 and Degree 4 MASS, there will be no crew
on board to repair traditional machinery problems
should they occur. Losses will increase in the event of
vessel collisions, vessel groundings, fires, and
explosions due to traditional machinery problems,
and even more so in the event of flooding [40]. The
next section researches the challenges faced by MASS
in applying COLREGs to vessel manoeuvring
practices from three perspectives: (1) MASS's
understanding of COLREGs in vessel manoeuvring
practices; (2) the intelligent ethical issues that may
arise for MASS in vessel manoeuvring practices; and
(3) MASS's intelligent ethical challenges to COLREGs.
2.1 MASS's understanding of COLREGs in vessel
manoeuvring practices
Although vessels like the Yara Birkeland [35] may
serve as a powerful indication for a shift in the
shipbuilding industry, their design is better suited for
coastal port waters [23]. However, more than 50
percent of vessel safety accidents occur in coastal port
waters [12]. At the same time, MASS has the highest
navigational risk of vessel collisions when navigating
in coastal port waters [7]. Especially in the case of
alternating humans and intelligents in MASS (Degree
3), the risk of negative consequences is increased [37].
Then, MASS needs to pay more attention to
navigational safety in coastal port waters and follow
the “ordinary practice of seamen” to fulfil the
COLREGs to ensure the safety of the vessel's
navigation.
The direct reference for the navigational safety of
MASS is the traditional vessel. If MASS is to be used
as a replacement for traditional vessels, the
navigational safety requirements for MASS are higher
than those for traditional vessels [32], and the
minimum requirements should also be as safe as for
traditional vessels [36]. Any collision avoidance and
preventive measures that can be achieved by
traditional vessels in vessel manoeuvring practice
should also be achieved by MASS. The design of the
vessel should also be consistent with the navigational
safety requirements of traditional vessels. In response
to the discussion of MASS navigational safety issues,
many academics have emerged from the collision
avoidance algorithms for MASS [3] [5] [38], MASS
applies to the provisions on flag state jurisdiction and
COLREGs in UNCLOS [9] [20], MASS security issues
[26], and the cybersecurity issues of MASS [29].
However, few academics have discussed the
application of MASS to COLREGs from the
perspective of MASS vessel manoeuvring practices.
Why discuss it from the perspective of vessel
manoeuvring practices? The reason is that COLREGs
involve both legal and technical issues of vessel
manoeuvring practices. Therefore, the interpretation
of COLREGs should be interpreted from the
perspective of legal norms as well as from the
perspective of technical norms of vessel manoeuvring
practices. When interpreting COLREGs from the
perspective of legal norms, it is necessary to strictly
follow the concepts, meanings, logics, and contextual
links stipulated in the rules to interpret the terms;
577
when interpreting COLREGs from the perspective of
technical norms of vessel manoeuvring practice, it is
necessary to follow the natural laws of collision
avoidance (such as the geometric principle of collision
avoidance) and the “good seamanship” summed up
by the seamen based on many years of practical
experience to interpret the specific collision avoidance
principles and avoidance actions [34]. At the same
time, legal interpretations and technical specifications
of vessel manoeuvring practices cannot be distinctly
separated or confused [41].
Example: COLREGs Rule 5: Every vessel shall at
all times maintain a proper look-out by sight and
hearing as well as by all available means appropriate
in the prevailing circumstances and condition so as to
make a full appraisal of the situation and of the risk
collision.
Some academics are of the view that, from the
point of view of legal interpretation, the above-
mentioned rule on “look-out” only mentions “every
vessel” and does not explicitly state that the
responsibility of lookout is the responsibility of “the
master and crew” [20]. Thus, COLREGs seem to be
applicable to MASS. However, the requirement to
maintain a regular lookout is reflected in “good
seamanship” in the Convention on the International
Regulations for Preventing Collisions at Sea, 1960, and
earlier regulations. This states that it does not exempt
the master or crew from liability for the consequences
of any negligence in maintaining a regular lookout
[34]. Thus, the master or crew is assigned lookout
duty under the COLREGs. Simultaneously, the term
“vessel” in this rule can be interpreted, from a
prudent and reasonable standpoint, to mean not only
the objectivity of the “vessel” itself but also all
available seamen on board, navigational aids, as well
as any appurtenances and installations on board, and
any manner and means.
From the point of view of vessel manoeuvring
practice, the lookout approach is one in which all
effective means can be employed with the aim of
continuously monitoring nearby vessels. In vessel
manoeuvring practice, the officer on watch (OOW)
maintains an uninterrupted lookout because the
seaman is required to determine in real time the
dynamics of any of the surrounding vessels, even if
they are only at potential risk of collision with the
vessel or are not at risk of collision for the time being.
In the practice of vessel manoeuvring, vessel “A,”
which is not in danger of collision with our vessel,
vessel “A,” may take measures to adjust its course or
speed to avoid collision with vessel “B,” whereas
vessel “A,” which has adjusted its course and speed,
may be at risk of collision with our vessel. Therefore,
from the perspective of vessel manoeuvring practice,
it is important to pay attention not only to the simple
head-on situations of crossing and overtaking but also
to the motion situations of all the vessels around our
vessel.
In heavy coastal waters, a vessel may need to
consider collision avoidance with more than one
vessel at a time. However, if one of these vessels
changes course or speed, the OOWs will need to
immediately assess the new situation and make a new
avoidance strategy. As well, when the vessel is sailing
in restricted waters or has poor visibility in a coastal
harbour, the master will assign the bosun to the bow
to assist the OOWs in keeping a lookout and
preparing to drop the anchor. The two most basic
ways to keep lookout are usually with the naked eye
and radar. The use of naked-eye observation can be
seriously affected in conditions of low visibility, and
there are inherent defects in radar, including the
inherent “blind zone” defects of radar, and the fact
that wind, waves, rain, and snow can interfere with
the use of radar or make it unusable when it is
interfered with by other military equipment in coastal
areas. Therefore, there may be errors in the radar-
scanned echoes, and sometimes the radar-scanned
object echoes are difficult to use as collision avoidance
by vessels, which requires the experience of the
OOWs to judge them. Additionally, the weather at sea
is changeable, so in order to enable OOWs to have a
clearer lookout, wipers and glass heating appliances
will be added to the glass of the vessel's bridge.
Furthermore, in the actual navigation of the vessel,
radar interference and poor visibility may occur at the
same time. In this situation, the vessel cannot
immed
iately stop and drop anchor to wait for the
weather to clear, as the actual depth of the sea and the
restricted area do not allow dropping anchor. After
that, the master will continue to command the vessel
on the bridge with his physical perception of the
environment as well as his extensive navigational
experience. Nevertheless, MASS (Degree 3) shore-
based remote operators are not able to be in a real
environment or be aware of their environment. It is
also important to consider the navigational aids and
auxiliary appliances required in vessel manoeuvring
practices to cope with the various special weather
conditions at sea. MASS (Class 3) should not be
considered to be directly applicable to COLREGs if
MASS cannot fulfil the minimum requirements for
traditional vessels.
As mentioned in this paper, OOWs require
monitoring the motion situation of the nearby vessels
uninterruptedly, as it is necessary to assess the
intention of other vessels to change their motion
situation, which is also the basis for subsequent
reasonable collision avoidance measures. In the case
of MASS (Level 3) man-machine coexistence, if
intelligence is involved in the pre-collision avoidance
measures or during the collision avoidance measures,
the transition between man and machine requires a
long period of time to judge and measure the
avoidance situation, which is a serious threat to
navigational safety and should not be regarded as a
full, uninterrupted, real-time performance of lookout
duties. Because COLREGs are both technical and legal
norms, it takes collaborative research by experts in
both maritime law and nautical technology to achieve
the natural unity of the two [26]. Therefore, the
applicability of MASS to COLREGs cannot be judged
and interpreted solely from the legal point of view but
also from the technical point of view of navigational
practice.
2.2 Intelligent Ethical Dilemmas of MASS in
Manoeuvring Practice
The IMO defines MASS (Degree 4) as a fully
autonomous, intelligent vessel [16]. An artificial
intelligence application needs to take into account its
578
operating environment in order to function properly
[28]. In coastal harbour-restricted waters, the width of
the channel becomes narrower, the depth of water
becomes shallower, and the distance from the shore of
the vessel becomes closer. As a result of these
conditions, the speed of the vessel also needs to be
higher, because a reduction in speed affects the
rudder effect or even leads to the disappearance of the
rudder effect [39]. The vessel sails in a narrow
channel, its manoeuvrability is somewhat restricted,
and it cannot steady course if the rudder effect
diminishes or fails [8]. Vessel collisions frequently
occur in narrow channels due to certain
characteristics, such as variable width, water depth,
and heavy waters in coastal ports. This severely
hinders the development of marine transportation
[13] [21]. When both channels are crowded and there
is a high vessel density in the narrow channel along
the coast, there is very little chance that a vessel will
alter its course [15].
In coastal harbours, there are often many small
vessels in narrow channels, and often it is impossible
to determine whether these vessels are fishing (Rule
3(d): The term “vessel engaged in fishing” means any
vessel fishing with nets, lines, trawls, or other fishing
apparatus that restricts manoeuvrability, but does not
include a vessel fishing with trolling lines or other
fishing apparatus that does not restrict
manoeuvrability), as well as fishing vessels made of
wood. Furthermore, a significant risk to MASS is
posed by these wooden fishing vessels with small
object markers, weak radar echoes, small vessels with
poor communications, and fishing vessels that are
unable to recognize the proper action [6]. Because
most small coastal vessels do not understand or pay
attention to COLREGs and information
communication, it is challenging to use collision
avoidance prediction models or algorithms for
collision avoidance actions. This makes it difficult for
large cargo vessels to coordinate avoidance with
them, and it also results from the unstable course and
speed of small vessels. Small fishing vessels are
unfocused and prone to abrupt course changes,
abrupt accelerations, abrupt decelerations, and abrupt
stops. If OOWs want to “departure from these rules”
and avoid the “immediate danger” as defined by the
COLREGs, they must therefore pay attention, avoid
collisions as soon as possible, and act quickly to
confirm and communicate with other large cargo
vessels in the area.
Furthermore, systems that deploy artificial
intelligence in the field will eventually have to make a
decision between two potentially undesirable
outcomes [28]. For example, (1) when a collision
between MASS and either of the other two vessels is
unavoidable, MASS needs to make a choice of which
vessel to collide with; (2) if MASS is restricted from
manoeuvring in restricted waters in narrow channels
along the coast, for example, when encountering
small vessels in narrow channels that do not comply
with COLREGs, fishing vessels that do not recognize
the correct action, or wooden fishing vessels that do
not have the ability to communicate information,
MASS has to make a choice of whether it will keep its
speed and its course, or will it steer out of the way.
In response to the above situation of MASS
choosing between two potentially unfavourable
outcomes, this paper proposes the following
questionnaire for seamen in the unlimited area of
navigation with different duty, vessel types, and
backgrounds:
Table 1. Basic information about the questionnaire
respondents)
________________________________________________
Category Subcate- Quan- Other Careers Vessel Type
gory tities
________________________________________________
A A1 5 - oil and chemical
(Master) A2 1 maritime oil and chemical
superintendent
A3 2 - bulk and container
A4 1 university bulk and container
professor
A5 1 Ph.D. and bulk and container
maritime lawyer
B B1 3 - oil and chemical
(Chief Mate) B2 2 - bulk and container
C C1 5 - oil and chemical
(Second or C2 5 - bulk and container
Third mate)
D D-B2 1 pilot I oil and chemical
(Pilot) D-C2 1 pilot II bulk and container
E E-C1 1 maritime oil and chemical
(Maritime lawyer
lawyer) E-C2 1 maritime bulk and container
lawyer
F F-C1 1 MSA official oil and chemical
(MSA) F-C2 1 MSA official bulk and container
________________________________________________
The respondents in Table 1 gave different answers
from their own backgrounds, as follows:
Table 2. Different backgrounds lead to different choices of
collision avoidance measures
________________________________________________
Category Possible measures for collision avoidance
________________________________________________
A1 For instance, in extremely harsh circumstances, a VLCC
(Very Large Cargo Carrier—Crude Oil Tanker) master
will ram any vessel that gets in the way with great
directness. Additionally, the master clarified that a fully
loaded VLCC can carry up to 300,000 tons of cargo oil,
that the draft may exceed 21 meters, and that the vessel's
manoeuvrability is very poor, making it difficult to steady
course and slow down speed in order to avoid the vessel.
There is a chance that oil will spill if it runs aground.
A2 Make the avoidance limit comprehensively based on the
vessel type's characteristics, report to VTS for
coordination, make full use of the vessel's manoeuvring
limits, pass as close to the obstructing vessel as possible,
and avoid using a large rudder angle in the last stage of
avoidance. Either the anchor is used before going
aground, or the engines and rudders cooperate. If there is
only an extreme choice, the choice will be to avoid the
obstructing vessel, and human life will be the first
priority.
A3 Evaluate the limits of the vessel's manoeuvring and, if
only extreme choices are available, will not choose to run
aground and will keep the vessel's speed to a minimum to
minimize the damage caused by a collision.
A4 If only extreme choices can be made, actively choose to
run aground, with human life being the priority.
A5 It's hard to answer, and it’s a test not of vessel
manoeuvring skills but of human nature.
B1 and Questionnaires were sent to a total of five people, four of
B2 whom did not respond. Another considered that, if left
with an extreme choice, it would be preferable to run
aground rather than to injure human lives.
C1 and The answer is pretty much unanimous: comply with the
C2 COLREGs and prefer to run aground rather than collide
with the vessel.
D-B2 Reduce the vessel's speed to the minimum speed that can
and be maneuvered, and drop anchor if necessary. At the
D-C2 same time, contact VTS for assistance and send tugs and
pilot boats to drive away the obstructing vessel.
E-C1 Comply with COLREGs and reduce the speed of the
vessel, but reducing the speed dramatically in a narrow
channel does not seem to be “good seamanship.”
However, in any event, even a collision cannot be deemed
to be subjectively intentional or reckless, as this would
579
affect the application of the “exemption from liability”
provisions of the Maritime Law.
E-C2 Even though the waters of the narrow channel are
restricted and such situations have been encountered
before, he is confident that he can pass through safely. At
the same time, it was felt that it took courage to navigate
the vessel to avoid the collision and that it was necessary
to be both careful and courageous.
F-C1 and The master was immediately called to the bridge, and
F-C2 until the master had clearly handed over command,
collision avoidance measures would be strictly enforced.
At the same time, it was considered that the consequences
of a collision could lead to the sinking of the obstructing
vessel or casualties, which could be more serious than a
grounding.
________________________________________________
Based on the information from the questionnaires
in Tables 1 and 2 above, it can be seen that
phenomena such as virtue, humanity, courage,
subjectivity, value judgments, and different types of
vessels leading to different ways of manoeuvring are
likely to be involved in extreme vessel manoeuvring
for collision avoidance. Realizing virtue in artificial
intelligence applications is inherently challenging. For
example, how do you quantify a brave algorithm
through a rule-based approach? One view that
contrasts with virtue is the utilitarian view, which
holds that ethical decisions maximize value [10].
Assuming that MASS can be as fully autonomous as
humans and has the ability to learn on its own, what
choice will MASS make in the face of the possibility of
causing danger to human lives or causing significant
environmental pollution and economic loss? In this
way, the intelligent ethical dilemma of MASS arises.
2.3 The MASS Intelligence Ethical Dilemma for
COLREGs
Moral decision-making cannot be explained solely
through moral dilemmas but is a situation-dependent
process of autonomous agent behaviour [2]. The
navigational environment of MASS in narrow coastal
channels is more complex and scenario-variable than
that of self-driving cars. Some academics have already
researched the moral impact of self-driving cars in
terms of the perception that self-driving car
algorithms will ultimately have to make decisions that
negatively impact passengers or other road users and
that moral judgments about this decision will have
implications for both self-driving car algorithms and
policy implementation [11] [18]. Then, the decisions
MASS makes when faced with intelligent ethical
dilemmas will also pose a challenge to COLREGs.
According to Rule 18 of the COLREGs,
“Responsibility between vessels,” the motor vessel
shall give way to the vessel engaged in fishing. Under
rule 9 of the COLREGs, “narrow channels,” a vessel
engaged in fishing shall not impede the passage of
any other vessel navigating within a narrow channel
or fairway. It follows that a vessel engaged in fishing
should not impede a MASS navigating in a narrow
channel or fairway, and the emphasis here is only on
“should not impede,” without altering the duty to
give way to the motoring vessel normally navigating
in a narrow channel.
If COLREGs apply to MASS, MASS is still
responsible for giving way to fishing vessels. As a
general guideline, an attempt should be made to
achieve a DCPA (distance closest point of approach)
of 2 nautical miles on the high seas and 1 nautical mile
on restricted waters [22]. A channel of 2 nautical miles
in width is usually regarded as a narrow channel [34].
And within the narrow channel, it is difficult for
MASS to choose to pass at the safest meeting distance
(1 nautical mile). If the MASS turns at a large degree
of course or reduces speed significantly to avoid an
obstructing vessel or a vessel with a risk of collision in
restricted waters, does the MASS need to take into
account the grounding factor? If MASS considers the
grounding or reefing factor, does MASS have the
courage and confidence to pass close to obstructing
fishing vessels? And if MASS is brave and
courageous, then MASS has a courageous “risk-
taking” attitude towards the possibility of collisions
that could result in injury or death?
Does a MASS have to follow the COLREGs by
changing course or reducing speed to avoid a collision
if a small vessel, a fishing vessel that is incapable of
recognizing the proper action, a wooden fishing
vessel that is unable to communicate, etc., not only
obstructs the MASS while it is navigating normally in
a narrow channel but also creates a risk of collision?
This is because the goal of the COLREGs is to prevent
vessel collisions. Is the MASS obligated to comply
with COLREGs by changing course or reducing speed
to avoid collisions, given that the goal of COLREGs is
to prevent vessel collisions? But MASS may seriously
harm the environment or result in economic damage
if it conforms with the COLREGs; if not, it violates the
regulations and runs the risk of causing vessel
accidents that threaten human life.
In summary, the challenges facing COLREGs are:
(1) When presented with a decision between perhaps
inflicting considerable property damage and
potentially endangering human life, how should
MASS make it? (2) The understanding and application
of COLREGs vary from person to person and context
to context, so there may be a judgment of subjective
thinking in their application. So how can the
subjective thoughts of MASS be judged?
3 RESULTS AND DISCUSSION
The navigational safety of MASS is largely based on
assumptions about the applicability of COLREGs. The
need for MASS to fulfil all the technical requirements
and devices of traditional vessel manoeuvring
practices and the ethical dilemmas of MASS may
seriously hamper the development of this new
technology. COLREGs are not set in stone but are a
living instrument. MASS can apply to COLREGs only
if they satisfy the requirements of the same
performance as traditional vessel manoeuvring and if
they resolve ethical dilemmas. In order to promote
MASS operations and navigational safety, the
following issues will be discussed: (1) Is it feasible to
revise the COLREGs? (2) Is it feasible to expand the
pilotage distances in conjunction with marine traffic
management?
580
3.1 Revision of the feasibility of COLREGs
The “look-out” rule of the COLREGs has attracted
much attention, and some academics have suggested
that it should be amended to allow look-out by
“computer vision” only [44]. However, the COLREGs
regulations are intrinsically related to each other and
do not stand alone. The COLREGs should be viewed
as a whole and should not be discussed in isolation. It
should not be assumed that once MASS fulfils the
requirements of a certain regulation, it can be directly
concluded that MASS can apply COLREGs in a
comprehensive manner.
COLREGs Rule 2 (a): ordinary practice seamen:
There is no rigidity of text as to what the usual
practice of seamen means; it is a question of
ascertaining its facts in the light of all the relevant
circumstances [24]. Furthermore, good seamanship
means good practice, while ordinary practice seamen
means general practice and alludes to good general
practice [34]. Therefore, such precautions as may be
required by the ordinary practice of seamen or by the
peculiar circumstances of the time are basically
regarded as an expression of good seamanship. So,
“ordinary practice seamen” is correlated with “good
seamanship.” COLREGs rule 5 “look-out”: Seamen, in
carrying out their lookout duties, should apply all
information obtained from the lookout to the vessel's
collision avoidance using good seamanship. So, “look-
out” is correlated with “good seamanship.” The “safe
speed” rule referred to in Rule 6 of the COLREGs:
Safe speed needs to be based on the navigational
environment and navigational waters of the formal
lookout. The formal lookout includes the density of
vessels in the nearby sea area, visibility, wind and
wave flow conditions, other vessel encounter
situations, and so on. After a comprehensive
judgment of the adoption of a safe speed, the lookout
is also a means of judging whether there is a risk of
collision with other vessels. So, “look-out” is
correlated with “safe speed.” The “risk of collision”
rule referred to in Rule 7 of the COLREGs: The
reference to radar and radar plotting in paragraph (b)
of this article is to the navigational aids and collision
avoidance calculations contained in the “look-out”
rule, and in the process of collision avoidance
calculations, a safe speed is to be used. In vessel
manoeuvring practice, most vessels sail at a constant
speed in order to allow the other vessel to measure
TCPA and DCPA and to use good seamanship to
avoid the “close-quarters situation” in COLREGs Rule
8. The COLREGs Rule 2 (b) “departure from these
rules” to avoid “immediate danger” is often faced in
restricted waters with dense traffic in coastal ports.
This is because in congested waters with vessel traffic,
such as harbours and narrow channels, where
multiple vessels can meet and pose a danger of
collision, there are special circumstances in which
deviations from the rules may be made. Similarly,
“ordinary practice of seamen” requires the application
of "good seamanship" and conforms to “ordinary
practice of seamen in special circumstances.”
In summary, the “look-out” rules can essentially
establish a direct or indirect link with any other rules.
There is also an intrinsic link between the various
rules of the COLREGs. The discussion of the
COLREGs' rules cannot be analysed separately. If
MASS applies COLREGs, then any of the
requirements of the COLREGs should be applied
without any conflict, and MASS needs to meet the
same navigational safety capabilities, conditions,
facilities, and appliances as a traditional vessel. If the
COLREGs are revised, the conflict between and the
application of MASS to any of the COLREGs' rules
will need to be considered. Avoiding conflicts of
understanding or difficulties of application between
traditional vessels and MASS. This is because these
conflicts of application may increase the number of
marine traffic accidents in the restricted waters of
coastal ports.
3.2 Maritime traffic management in conjunction with
extended pilotage distances
Maritime traffic management can be categorized into
VTS (Vessel Traffic Service), Vessel Reporting System,
and Vessel Routing System [17]. International
Convention for the Safety of Life at Sea (SOLAS)
regulations V/11(1) and V/12(1) provide that the
objectives of both vessel reporting and VTS are to
assist vessels in their efforts to ensure the safety of life
at sea, the safety and efficiency of navigation, and the
protection of the marine environment. Under
regulation 12(2) of the SOLAS, states have the right to
establish VTS when they consider that the volume of
traffic or the level of risk warrants the provision of
such services, and the IMO Guidelines and Principles
on Vessel Reporting and VTS imply that these
systems are to be applied by the “seamen” or the
“master” [4] [32]. At the same time, under SOLAS
regulation V/10, the coastal state has the right to
impose a mandatory vessel routing regime on foreign
vessels, and there is a correlation between the TSS and
COLREGs rule 10 traffic separation scheme. Then, in
order to address the possible ethical challenges and
navigational safety of MASS without modifying the
COLREGs, it would be more appropriate for MASS to
be piloted by human beings in coastal harbour waters.
From the previous questionnaire survey, it can be
seen that the measures available to the pilot to avoid
vessel collisions are the most numerous and relatively
adequate, and he can always deploy tugs, pilot boats,
and all the means of the maritime traffic management
system to avoid the occurrence of accidents, so the
extension of the pilotage distance is a more
appropriate option. COLREGS could not be
successfully used to regulate traffic flow, so another
solution was needed. This solution takes the form of
TSSs in areas with high concentrations of traffic [30].
The specific location of the pilotage station and the
specific pilotage distance can be set by each country
through a comprehensive assessment of the traffic
density and the frequency of traffic accidents. If a
traffic accident occurs in the course of piloting, the
issue of liability between the pilot and MASS can also
be determined in accordance with each country's
domestic pilotage-related laws.
4 CONCLUSION
The continued development of MASS in commercial
operations requires compliance with COLREGs for
navigational safety, but it faces challenges: (1) The
581
question of whether MASS satisfies the application of
COLREGs in vessel manoeuvring practice; (2) MASS
may present an ethical question of choosing between
a threat to human life or a major environmental
contamination or damage to property.
According to the research mentioned earlier in this
paper, the MASS shall satisfy all means of collision
avoidance and appurtenances that are consistent with
traditional vessel manoeuvring practices to ensure
that the MASS is applicable to COLREGs. At the same
time, a questionnaire survey was conducted among
respondents of different backgrounds and ranks. It is
concluded: (1) The attitude of the master of a large
dangerous goods vessel, in extreme and unavoidable
circumstances, is to prefer a collision to running the
vessel aground, even though a vessel collision will
probably result in loss of life or injury; (2) The attitude
of masters of general cargo and container vessels, in
extreme and unavoidable circumstances, is to reduce
speed as much as possible, slowing down with an
anchor if necessary, to minimize collisions and to
reduce damage; (3) The chief, second, and third mates,
usually choosing human life first, will actively choose
to run aground; (4) The pilot's attitude is to use the
vessel's manoeuvring to the utmost and to cooperate
with all measures, such as anchors, VTS, tugs, pilot
boats, etc., to drive away obstructing vessels; (5) The
maritime lawyer's attitude was that he would not
make any choices that might increase his legal liability
and that even if a collision might occur, he had not
acted deliberately or recklessly because of the
exemption from liability in the Maritime Law; (6) The
attitude of the officials of the MSA is that they are
more concerned with personal injuries and deaths at
sea than with marine traffic accidents. Assuming then
that MASS is autonomous and capable of self-
learning, it will face ethical choices, and artificial
intelligence, which is primarily concerned with
helping and serving humans, must not be given
choices that could harm human life. If MASS does not
have enough courage and bravery, then it is likely to
increase the rate of marine traffic accidents when
sailing in restricted coastal waters.
Due to the dual combination of legal and technical
attributes of COLREGs, there are different
interpretations and applications of COLREGs in
different contexts. Therefore, the application of MASS
to COLREGs cannot be judged solely on the basis of
the interpretative standards of the law. If the option to
revise COLREGs to apply MASS is selected, then the
systemic nature of COLREGs should be taken into
account to avoid conflicts of understanding and errors
of application when used simultaneously with
traditional vessels.
Finally, this paper argues that COLREGs are
subjective and value-judging, so they are applicable to
vessels piloted by human beings, and suggests that
the IMO should be prudent in applying COLREGs
directly to MASS before comprehensively considering
MASS's navigational safety.
REFERENCES
[1] Akdag M, Solnor P, Johansen,TA. 2022. Collaborative
collision avoidance for Maritime Autonomous Surface
vessels: A review. OCEAN ENGINEERING, 250.
(DOI:10.1016/j.oceaneng.2022.110920) .
[2] Amit K.C, Munindar P.S. 2018. Sociotechnical Systems
and Ethics in the Large. In Proceedings of the 2018
AAAI/ACM Conference on AI, Ethics, and Society (AIES
’18), 48–53. (DOI:10.1145/3278721.3278740).
[3] Anete V, Bye, Robid O, Robin T.B, Ottar L.O, Thor I.F.
2021. Path planning and collision avoidance for
autonomous surface vehicles I: a review. Journal of
Marine Science and Technology, 26, 1292-1306.
(DOI:10.1007/s00773-020-00787-6).
[4] ANNEX 10. RESOLUTION MSC.428(98). MARITIME
CYBER RISK MANAGEMENT IN SAFETY
MANAGEMENT SYSTEMS. Adopted on 16 June 2017.
https://wwwcdn.imo.org/localresources/en/OurWork/Se
curity/Documents/Resolution%20MSC.428(98).pdf
(access May 12, 2024).
[5] Bjrn-Olav H.E, Morten, B Wilthil, Erik F.W, Andreas L.F,
Edmund F.B. 2019. The branching-course model
predictive control algorithm for maritime collision
avoidance. Journal of Field Robotics, 36 (7), 1222–1249.
(DOI:10.1002/rob.21900).
[6] Chia-Hsun C, Christos K, qing Y, Zaili Y. 2021. Risk
assessment of the operations of maritime autonomous
surface vessels. Reliability Engineering & System Safety,
207. (DOI:10.1016/j.ress.2020.107324).
[7] Chien C.C, Chia N.W, Hsien P.H. 2022. A novel
quantitative and qualitative model for forecasting the
navigational risks of Maritime Autonomous Surface
vessels. Ocean Engineering, 248,110852.
(DOI:10.1016/j.oceaneng.2022.110852).
[8] Dam K.T, Tanimoto K, Fatimah E. 2008. Investigation of
vessel waves in a narrow channel. Journal of Marine
Science & Technology. 13, 223-230. (DOI:10.1007/s00773-
008-0005-6).
[9] Dong B.Y, Lowell B, Zhu L. 2024. Navigating uncharted
waters: Challenges and regulatory solutions for flag
state jurisdiction of Maritime Autonomous Surface
vessels under UNCLOS. Marine Policy, 161, 106039.
(DOI:10.1016/j.marpol.2024.106039).
[10] Don H, Ioan M. 2017. Artificial Moral Cognition: Moral
Functionalism and Autonomous Moral Agency.
Springer International Publishing, Cham, 121–159.
(DOI:10.1007/978-3-319-61043-6_7).
[11] Edmond A, Sohan D, Richard K, Jonathan S, Joseph H,
Azim S, Jean-François B, Iyad R. 2018. The Moral
Machine experiment. Nature, 563, 59-64. (DOI:10.1038/
s41586-018-0637-6).
[12] EMSA marine casualties and incidents 2023 annual
overview. https://www.iims.org.uk/emsa-marine-
casualties-and-incidents-2023-annual-
overview/#:~:text=EMSA%20marine%20casualties%20an
d%20incidents%202023%20annual%20overview,Incident
s%20of%20pollution%20...%204%20Safety%20investigati
ons%20 (access May 17, 2024).
[13] Gao X, Makino H, Furusho M. 2017. Analysis of vessel
drifting in a narrow channel using Automatic
Identification System (AIS) data. WMU Journal of
Maritime Affairs. 27, 351-363. (DOI:10.1007/s13437-016-
0115-7).
[14] Hans v.d.B, Jaco G, Monique V.D.D. 2020. Meaningful
Human Control in Autonomous vesselping: An
Overview. IOP Conference Series Materials Science and
Engineering. 929(1):012008. (DOI:10.1088/1757-
899X/929/1/012008).
[15] Hu H.T, Chen X.Z, Sun Z. 2017. EFFECT OF WATER
FLOWS ON vessel TRAFFIC IN NARROW WATER
CHANNELS BASED ON CELLULAR AUTOMATA.
Polish Maritime Research, 24(s1). (DOI:10.1515/pomr-
2017-0115).
[16] IMO. IMO’s Maritime Safety Committee finalizes its
analysis of vessel safety treaties, to assess next step for
regulating Maritime Autonomous Surface vessel
(MASS).
582
https://www.imo.org/en/MediaCentre/PressBriefings/pa
ges/MASSRSE2021.aspx (access May 20, 2024).
[17] Iva P. 2023. Can the International Regulatory
Framework on vessels’ Routing, vessel Reporting, and
Vessel Traffic Service (VTS) Accommodate Marine
Autonomous Surface vessels (MASS)? Ocean
Development & International Law, 54(2).
(DOI:10.1080/00908320.2023.2211781).
[18] Jean-François B, Azim S, Iyad R. 2016. The social
dilemma of autonomous vehicles. Science, 352, 6293,
1573–1576. (DOI:10.1126/science.aaf2654).
[19] Jialun L, Muhammet A, Emre A, Ozcan A, Esma U,
Rafet E, Osman T. 2022. Prediction of human-machine
interface (HMI) operational errors for maritime
autonomous surface vessel (MASS). 27, 293-306.
(DOI:10.1007/s00773-021-00834-w).
[20] Justyna N, Zuzanna P.D. 2019. Revolution or Evolution?
Challenges posed by autonomous vessels for national
and international legal framework, Comp. Law Rev. 25,
239–255. (DOI:10.12775/ CLR.2019.008).
[21] Lee C.K, Moon S.B, Jeong T.G. 2016. The investigation
of vessel manoeuvering with hydrodynamic effects
between vessels in curved narrow channel. International
Journal of Naval Architecture & Ocean Engineering,
8(1), 102-109. (DOI:10.1016/j.ijnaoe.2016.01.002).
[22] Lee, G.W.U, Parker, J. 2007. Managing Collision
Avoidance at Sea. London: Nautical Institute.
[23] Maersk CEO: Unmanned Container vessels Not in My
Lifetime. 2018. OFFSHORE ENERGY.
https://www.offshore-energy.biz/maersk-ceo-
unmanned-containervessels-not-in-my-lifetime/ (access
May 15, 2024).
[24] Mankabady S. 1978. Collision at Sea, A Guide to the
Legal Consequences. North-Holland Publishing
Company, P.91.
[25] Miso M, Fausto F. 2022. Autonomous Surface Vessels
and COLREGs: Considering the Amendments. Pored.
Pomor. Pravo, 61(176), 323–362.
(DOI:10.21857/m3v76t5kdy).
[26] Mohamad I, Adrian I, Hussein I, Patrick R. 2022.
Maritime Autonomous Surface vessels: Problems and
Challenges Facing the Regulatory Process.
Sustainability, 14(23). (DOI:10.3390/su142315630).
[27] Mustaffa M, Munawwarah A, Shaharudin A, Nazli
A.A. 2017. Marine Traffic Density Over Port Klang,
Malaysia Using Statistical Analysis of AIS Data: A
Preliminary Study. Journal of ETA Maritime Science,
4(4):333-341. (DOI:10.5505/jems.2016.60352).
[28] Niels v.B, Benjamin T, Jorge G, Simo H. 2020. Human-
Centred Artificial Intelligence: A Contextual Morality
Perspective. Behaviour and Information Technology,
41(2). (DOI:10.1080/0144929X.2020.1818828).
[29] Nimra T, Tsai C.L. 2024. Maritime Autonomous Surface
vessels: A Review of Cybersecurity Challenges,
Countermeasures, and Future Perspectives. IEEE, 99, 1-
1. (DOI:10.1109/ACCESS.2024.3357082).
[30] Phil B. 2002. A sociological interpretation of the
COLREGS. THE JOURNAL OF NAVIGATION, 55.
(DOI:10.1017/S0373463302001686).
[31] Resolution A.1158(32). GUIDELINES FOR VESSEL
TRAFFIC SERVICES. Adopted on 15 Dec. 2021.
https://wwwcdn.imo.org/localresources/en/KnowledgeC
entre/IndexofIMOResolutions/AssemblyDocuments/A.1
158%2832%29.pdf (access May 12, 2024).
[32] Rødseth Ø.J, Burmeister H.C. 2015. Risk assessment for
an unmanned merchant vessel. Int. J. Mar. Navig. Safety
Sea Trans. 9 (3), 357–364. (DOI: 10.12716/ 1001.09.03.08).
[33] Roman C, Tatiana Č. 2019. Impact of Digital
Technologies on Labor Market and the Transport Sector.
Transportation Research Procedia, 40, 994-1001.
(DOI:10.1016/j.trpro.2019.07.139).
[34] Si Y.Z, Wu Z.L. 1995. Collision Avoidance on vessels
Act，Dalian Maritime University Press.
[35] The first ever zero emission autonomous vessel.
Knowledge grows. https://www.yara.com/knowledge-
grows/game-changer-for-the-environment/ (access May
13, 2024).
[36] Thieme C.A, Utne I.B, Haugen S. 2018. Assessing vessel
risk model applicability to marine autonomous surface
vessels. Ocean Engineering 165, 140–154. (DOI:
10.1016/j.oceaneng.2018.07.040).
[37] Thomas T. 2019. Maritime Autonomous Surface vessels
(MASS) and the COLREGS: Do We Need Quantified
Rules Or Is “the Ordinary Practice of Seamen” Specific
Enough? Trans Nav the International Journal on Marine
Navigation and Safety of Sea Transportation, 13(3), 511-
518. (DOI:10.12716/1001.13.03.04).
[38] Tor A.J, Tristan P, Andrea C. 2016. vessel collision
avoidance and COLREGS compliance using simulation-
based control behavior selection with predictive hazard
assessment. IEEE Transactions on Intelligent
Transportation Systems, 17 (12), 3407–3422.
(DOI:10.1109/TITS.2016.2551780) .
[39] Wand L, Wu Q, Liu J.L, Li S.J, Rudy R.N. 2019. State-of-
the-Art Research on Motion Control of Maritime
Autonomous Surface vessels. Marine Science and
Engineering, 7(12), 438. (DOI:10.3390/jmse7120438).
[40] Wróbel K, Montewka J, Pentti K. 2017. Towards the
assessment of potential impact of unmanned vessels on
maritime transportation safety. Reliability Engineering
& System Safety. 165, 155–
169.(DOI:10.1016/j.ress.2017.03.029).
[41] Wu Z.L. 1990. On the Legal and Technical
Interpretation of the Collision Avoidance Rules at Sea.
Journal of Dalian Maritime Academy.
[42] Wu Z. L, Zhao Y. L. (2021). vessel Collision Avoidance
and Watchkeeping. Dalian: Dalian Maritime University
Press.
[43] Xue L, Kum F.Y. A human-centred review on maritime
autonomous surfaces vessels: impacts, responses, and
future direction. Transport Reviews.
(DOI:10.1080/01441647.2024.2325453) .
[44] Zhou X.Y, Huang J.J, Wang F.W, Wu Z.L, Liu Z.J. 2019.
A Study of the Application Barriers to the Use of
Autonomous vessels Posed by the Good seamen vessel
Requirement of COLREGs. The Journal OF
NAVIGATION, 73(3). (DOI:10.1017/S0373463319000924).
