Ocean Engineering

Volume 261, 1 October 2022, 112029

COLREGs and their application in collision avoidance

algorithms: A critical analysis
Jesús A. García Maza a , Reyes Poo Argüelles b

Show more

Outline Share Cite

https://doi.org/10.1016/j.oceaneng.2022.112029
Get rights and content

Under a Creative Commons license open access

Highlights
• Identification of basic criteria to consider when making decisions in vessel
encounters in accordance with COLREG.

• Identification of COLREGs' weaknesses/ambiguities, to unify patterns for the

development of collision avoidance algorithms.

• Identification of possible problems of interpretation and some common

mistakes found in recent or remarkable literature.

Abstract
Ship collisions are some of the biggest risks in shipping, and Decision Support System/Collision Avoidance-
Alert Systems (DSS/CAS) are being developed to prevent and avoid them. They must become an essential
equipment for any vessel, especially for Maritime Autonomous Surface Ships (MASS).

Assuming the Convention on the International Regulations for Preventing Collisions at Sea, 1972 (COLREG
72) as the basis for every DSS/CAS, there is growing concern in the maritime field about the lack of
consistency of published studies, in particular those on the design of algorithms for the control of ship
collision avoidance manoeuvres, with the requirements of the current COLREG.
The methods applied for the assessment of scenarios and situations will be analyzed, since only a correct
appraisal of the circumstances will bring about the adequate response for such critical situations.

COLREG is conceived to always have the Officers Of the Navigational Watch (OONW) (more specific than the
overly generic expression Officer On Watch -OOW), at the center of every decision.

Therefore, in the design of a DSS/CAS, interpretations that approach collision situations from a different
perspective than that of an OONW should be avoided. Safe communication with other vessels would
reinforce and improve the information available to the OONW.

This paper aims to offer an insight into ship collisions avoidance according to COLREG 72 that may prove
useful to the OONWs either on board or remotely, or even to autonomous systems. It is illustrated by
examples taken from some works which, despite their great influence in the current literature, do not have a
correct standpoint.

Previous Next

1. Introduction
A large number of studies point to human error as one of the prominent causes for maritime accidents
(ALLIANZ, 2020; Antão and Guedes Soares, 2008; EMSA, 2020; Wróbel, 2021).

Ship collisions rank high on the list of maritime accidents. Many efforts have been set up to prevent them
due to their harmful consequences and strategic importance on navigational safety (Zhang et al., 2018).

With the objectives of cost reduction, a race has begun towards the automation of ships aiming the
downsizing of humans and therefore of OONWs. The devising of DSS/CAS to ship's trajectory planning,
either as supporting elements or as direct decision making devices, has been one of the most common
proposals to date.

As the primary source to define the situations to be avoided, all DSS/CAS must count on COLREG 72, the
actual legal frame. An adequate knowledge of this standard seems crucial for the correct assessing of critical
situations and for taking adequate actions.

Literature about algorithms that can be effective in MASS is becoming extensive, with very diverse
approaches: fuzzy logic, evolutionary algorithms, neural networks, interval programming, 2D grid map,
collision threat parameters area techniques, the fast marching method, differential games, velocity obstacle
algorithms, etc.

However, when planning those algorithms in simulated or real scenarios, simplified and basic assumptions
of COLREG 72 are far from its real complexity. The oversimplification that is common in most formulations
reinforces the view of humans (OONW) as a weak link that can fail in an incomprehensible way facing a task
that, apparently, is very simple.

This paper aims to generate a reflection on what has to be taken into account in the design of the algorithms
that regulate the DSS/CAS or any equipment that attempts to perform a collision avoidance manoeuvre on a
ship. To this end, the current work will be developed in a way accessible to people unfamiliar with the
marine world.

The collision avoidance complexity can be assessed by showing the minimum requirements that a DSS/CAS
must meet according to the current legislation.

COLREG 72 Rules directly related to the evaluation of vessel encounters are analyzed, identifying possible
problems of interpretation and some common mistakes found in recent literature, in particular a possible
bad legacy from COLREG 60.

Finally, it should be noted that according to the definition of vessel given in COLREG 72 Rule 3:
“The word “vessel” includes every description of water craft, including non displacement craft, WIG craft
and seaplanes, used or capable of being used as a means of transportation on water.”

All are obliged to comply with COLREGs, irrespective of their size.

2. COLREGS analysis
As above mentioned and despite the fact that some authors highlight its complexity (Crosbie, 2009), its lack
of effectiveness (Belcher, 2002) and its unnaturalness (Kemp, 2009), the protocol to follow is COLREG 72. So,
a perfect knowledge of the current text is necessary to take the right decisions in these situations.

It starts by specifying the spatial scope of its application, as shown in Table 1.

Table 1. COLREG 72 spatial scope.

RULE DEFINITION

1 High seas and all waters connected therewith navigable by seagoing vessels.

9 Narrow channel or fairway

10 Traffic separation schemes

It should be noted that there may be different rules for certain navigation areas: (Suez Canal, 2015),
(Panama Canal, 2019), etc.

In no case is traffic density taken into account as a factor that generates actions other than those specified
for the indicated areas, beyond precautionary recommendations. Actions can be radically different from
place to place and this must be taken into account when designing an evasive manoeuvre.

Secondly, undesirable or avoidable situations have to be identified.

There is no definition of the term collision in COLREG 72.

After collision, close quarters is the second type of undesired event to consider. Wang et al. (2017) said:
“Nevertheless, there are no unified quantitative and qualitative interpretations on the term ‘close-quarters
situation’ in the world, and the exact definition is also not given in the COLREGs by far” (p.487). It is one of
the qualitative terms that need to be clarified (Tsai et al., 2017).
Rule 8 - Action to avoid collision, aims to describe safe passage between vessels. Although the information it
provides may not be categorical, it allows for a definition of close quarters as some distance between
collision and the safe distance of Rule 8 d), terms collected in Table 2.

Table 2. Types of encounters and their definitions.

ENCOUNTER RESULT RISK SITUATION

Collision Contact between vessels Critical

Close quarters Possible contact between vessels

Safe distance No contact between vessels Safe

Regarding the identification of critical situations, COLREG 72 Rule 7 b) and d), Risk of collision, states that the
OONW:
• Must use: RADAR plotting or equivalent systematic observation of detected objects.

• Must consider possible risk of collision: when the compass bearing of an approaching vessel does not
appreciably change, or even when an appreciable bearing change is evident, particularly when
approaching a very large vessel or a tow or when approaching a vessel at close range.

In order to do so, notice must be taken of: relative bearings (M°), and distances (d). Fig. 1 shows the plotting
used throughout this paper.

Download : Download high-res image (429KB)

Download : Download full-size image

Fig. 1. Ship plotting.

Since encounters are dynamic situations, continuous monitoring is required.

Thus, the distance at the Closest Point of Approach (CPA) allows us to identify the type of encounter as
collision, close quarters or safe distance and the Time of the Closest Point of Approach (TCPA) is used to give
priority to critical situations (Bole et al., 2014; Xu and Wang, 2014). CPA and TCPA terms are defined in IMO
(IMO, 1986a).

It is to be noted that both CPA and TCPA must be measured according to a reference, as the Consistent
Common Reference Point (CCRP) (IMO, 2004).

Measurement equipment is required to calculate distance. RADAR-ARPA is the only aid overtly
acknowledged by the COLREG 72, even though Rule 5 states that others could also be used. Along the same
lines IMO Resolution A.1106 (29) accepts the Automatic Identification System (AIS) for onboard use as
another useful tool to this end (IMO, 2015).

These calculations should start according to Rule 7 b): “… including long-range scanning to obtain early
warning of risk of collision …”

The term Long-Range Scanning (LRS) is not yet defined. For that, the difference in the distances at which the
involved vessels detect a collision risk situation can lead to confusing situations.

In the absence of RADAR or similar equipment (AIS), LRS is understood to be the visual or audible range. If
both RADAR and AIS exist, the first reliable information given by these devices should initiate the analysis of
the situation, even if the information comes from a long distance.

When a critical situation is observed, compliance with COLREG 72 is mandatory. Its rules are organized
around three major criteria: Visibility, Kind of Propulsion and Restrictions both technical and geographical,
discussed in COLREG sections 3 to 7, and outlined in Table 3.

Table 3. COLREG criteria.

VISIBILITY KIND OF PROPULSION RESTRICTIONS

Part A - General (Rules 1–3) Part A - General (Rules 1–3) Part A - General (Rules 1–3)

Part B-Steering and Sailing Part B-Steering and Sailing Part B-Steering and Sailing (Rules 4–19)

Section I Section I
Conduct of vessels in any condition of Conduct of vessels in any condition of visibility
visibility (Rules 4–8) (Rules 9 and 10)Technical

Section II Section II Section II

Conduct of vessels in sight of one Conduct of vessels in sight of one Conduct of vessels in sight of one another (Rule
another (Rules 11–17) another (Rule 18) 18) Geographical.

Section III
Conduct of vessels in restricted
visibility (Rule 19)

Let's take a deeper look at the COLREG criteria.

3. The visibility criterion in COLREG 72
Visibility is very often the only criterion taken into account when designing a DSS/CAS (Goerlandt and
Kujala, 2011; Lopez-Santander and Lawry, 2017; Montewka et al., 2010; Perera et al., 2009, 2010, 2011;
Smeaton and Coenen, 1990).

In addition to using this single criterion, it is usually simplified by reducing it to the vessels in sight scenario.
The reason for this may be that COLREG 60 continues to apply, as it includes nothing that prevents the
application of the rules used for vessels in sight to restricted visibility scenarios.

COLREG 72 is more complex and provides a customized response for each visibility scenario. Thus, Rule 3-
General Definitions, defines in section l) the term restricted visibility as “any condition in which visibility is
restricted by fog, mist, falling snow, heavy rainstorms, sandstorms or any other similar causes”. It is worth
mentioning that the literature has made a wide interpretation of “any other similar causes” as in Cockcroft
and Lameijer (2011): “Examples of ‘other similar causes’ are smoke from own vessel, other vessels, or
ashore, and dust storms” (p.16).

The conduct of vessels in restricted visibility or not in sight is governed by Rule 19 considering two possible
sequences of events described in sections d) and e).

So, in all, according to the visibility in the encounter area, COLREG 72 distinguishes between:
• Vessels in sight,

• Vessels not in sight with RADAR and

• Vessels not in sight without RADAR.

An analysis will be made of each scenario.

3.1. Vessels in sight

COLREG 72 Section II: Vessels in sight, considers three possible critical situations:
• Overtaking (Rule 13),

• Head-on (Rule14) and

• Crossing (Rule 15).

In the three situations, establishing the vessels’ Headings (and not Courses Over Ground (COGs)) is a must.
Once this is done, a priority system is established according to the Rules 16 and 17:
• Rule 16 - Give-way: keep out of the way of another

• Rule 17 - Stand-on: shall keep her course and speed (keep the way)

The system is further refined by Rules 8 f) iii), 17 a) ii), 17 b), which signal when a vessel the passage of
which is not to be impeded, or a stand-on vessel, must take an action to avoid collision. Therefore, it can be
referred to as a “limited priority system".
A brief description of each situation will follow, identified by the relative bearing and speed (Mo, SR), and
taking as reference heading and course of own and target ships (HO, CO, HT, CT), as shown in Table 4. The
values for Mo are deduced from the description of Rules 13, 14, 15, 19 and 21. Light sector tolerances defined
in Annex I. 9 of COLREG 72 are not considered for these values (at night, tolerances for the light sectors are:
0°/360° ± 3°; 112.5° and 247.5° ± 5°).

Table 4. Vessels in sight situations.

SITUATION Mo SR REFERENCE

VESSEL OVERTAKING 292.5° < Mo < 360° SR < SO HO , H T

0° ≤ Mo < 67.5°

VESSEL BEING OVERTAKEN 112.5° < Mo < 247.5° SR > 0 HO

HEAD-ON Not specified S R > SO HO , C O , H T , C T

CROSSING 247.5° ≤ Mo < 360° SR > 0 HO , H T

0° ≤ Mo ≤ 112.5°, as long as above situations have been discarded

3.1.1. Overtaking
Overtaking is the first situation to consider. It involves an Overtaking vessel and a Being Overtaken vessel, as
stated in Rule 13 a): “… any vessel overtaking any other shall keep out of the way of the vessel being
overtaken”.

The Being Overtaken vessel finds herself in an easier situation since, a priori, she needs neither to
manoeuvre nor to know the heading of the Overtaking vessel. By night, the lights of a vessel are a great help
as long as they are in visible range and a safe distance separates the vessels (Rule 22 - Visibility of Lights).

3.1.2. Head-on
Head-on is the hardest situation to determine. According to Rule 14 a), “When two power-driven vessels are
meeting on reciprocal or nearly reciprocal courses so as to involve risk of collision each shall alter her course
to starboard so that each shall pass on the port side of the other”.

The context of Rule 14 refers to heading (not course), mostly if section b) is considered:

Rule 14 b) Such a situation shall be deemed to exist when a vessel sees the other ahead or nearly ahead and
by night she could see the masthead lights of the other in line or nearly in a line and/or both sidelights and
by day she observes the corresponding aspect of the other vessel.

Phrases such as “ahead or nearly ahead” and “masthead lights of the other in line or nearly in a line” clearly
refer to the Headings of both vessels. In this case they are used as reinforcement in the description of the
situation, since course and heading are not the same. In fact, the IMO Sub-Committee On Safety Of
Navigation in the Standard Marine Communication Phrases (SMCP) (IMO, 2001) provides the following
definitions:
• Course: The intended direction of movement of a vessel through the water.
• Heading: The horizontal direction the vessel's bows at a given moment measured in degrees clockwise
from north.

In this regard, Bowditch (2019) warns: “It is easy to confuse heading and course. Heading constantly changes
as a vessel yaws back and forth across the course due to sea, wind and steering errors” (p.5).

Unfortunately, M° range is not specified in COLREG 72, although it was in past versions (Cockcroft and
Lameijer, 2011). The word nearly is difficult to quantify, and not without controversy.

3.1.3. Crossing
Crossing should be considered once Overtaking and Head-on situations have been discarded. Rule 15
describes the response to a crossing situation, using the vessel's heading as reference:
When two power-driven vessels are crossing so as to involve risk of collision, the vessel which has the other
on her own starboard side shall keep out of the way and shall, if the circumstances of the case admit, avoid
crossing ahead of the other vessel.

It is quite common in this situation to assume that, in terms of the give-way vessel changing course, the
reaction must be an alteration to starboard. However, for situations where the target vessel speed (ST) is
greater than own vessel speed (SO) and the vessels keep getting closer, a change to port may be more
effective and safer.

3.1.4. Rule 17, the limited priority

As already pointed out, in some situations in the vessels in sight scenario, a ship is given priority over
another. However, that priority is limited by Rule 17 b):
When, from any cause, the vessel required to keep her course and speed finds herself so close that collision
cannot be avoided by the action of the give-way vessel alone, she shall take such action as will best aid to
avoid collision.

The wording of this Rule may lead to errors of interpretation.

In the encounters between vessels each one knows her own manoeuvring characteristics, but doesn't know
those of the other vessel. With that perspective, the vessel that requires to keep course and speed might
have a problem deciding whether to act or not. And the decision is more likely to be wrong if the
manoeuvring characteristics of both vessels are quite different.

To avoid this added risk, it is necessary to establish the following requirement: the distance at which the
give-way vessel should start to manoeuvre must be greater than the distance set to comply with Rule 17 b)
for the stand-on vessel.

Only by knowing the safety distance defined for the stand-on vessel can it be respected, so that she is not
forced to manoeuvre.

This is a topic not usually addressed in the literature on collision avoidance, although it represents a
challenge in ship-to-ship encounters and suggests the need of sharing or communicating extra information
about the defined manoeuvring distances for the involved ships in the encounters.
More than ships' intentions (Du et al., 2020), information is what the OONW needs.

3.2. Vessels NOT in sight

Vessels not in sight scenario is key in the already mentioned oversimplification. COLREG 60 only mentions a
limitation in the manoeuvres to change course for navigation in restricted visibility:
6)” … An alteration to starboard, particularly when vessels are approaching apparently on opposite or nearly
opposite courses, is generally preferable to an alteration to port”.

An important change was introduced in COLREG 72, establishing rules to change course different to those
for vessels in sight. It has generated some controversy (Fujiwara et al., 2017; Salinas, 2006), and its rules are
difficult to be discerned by the OONWs (Mohovic et al., 2016). It must be clear for those in charge of the
navigation (humans or autonomous ships) which scenario are they in, to be sure that they apply the same
rules.

This scenario is subdivided in vessel not in sight with RADAR, and vessel not in sight without RADAR.

3.2.1. Vessels not in sight with radar

Few studies take into account vessels not in sight scenario and situations, or, when they do, they assume
conducts not included in COLREG 72 for a restricted visibility scenario (Hilgert and Baldauf, 1997;
Szlapczynski et al., 2018). There are no priorities in this scenario, so there is neither give-way nor stand-on
vessel.

There are three situations:

• Overtaking,

• Forward of the beam,

• Abeam or abaft the beam.

Overtaking appears in COLREG 72, Section III Rule 19 d) (i): “… other than for a vessel being overtaken”. But
the only reference to the overtaking concept is included in Section II: Conduct of Vessels in Sight of One
Another.

The problem of the possible extension to any state of visibility was clearly solved in COLREG 60, while a
rather obscure wording was approved in COLREG 72:
• COLREG 60 Rule 24 a) “Notwithstanding anything contained in these Rules, …”

• COLREG 72 Rule 13 a) “Notwithstanding anything contained in the Rules of Part B, Sections I and II, …”

Is the concept of overtaking limited to vessel in sight or could it be extended to any condition of visibility?
There is a problem of indeterminacy and therefore an increased risk of collision in the application of
COLREG Rule 19 d) when a vessel is being overtaken:
• the vessel being overtaken will manoeuvre, in application of Rule 19(d)(ii), avoiding an alteration of
course towards the vessel abaft the beam. It is not so clear how this (overtaking) vessel should
manoeuvre.
• It is not specified how will they manoeuvre when the vessel abaft the beam comes dead astern.

Any algorithm to be implemented in a CAS must be free from doubt and the interpretation should be
homogeneous for all CAS.

Forward of the beam, and abeam or abaft the beam characteristics are shown in Table 5.

Table 5. Vessels not in sight situations.

SITUATION Mo SR REFERENCE

FORWARD OF THE BEAM 270° < Mo < 360° SR > 0 HO

0° ≤ Mo < 90°

ABEAM OR ABAFT THE BEAM 90° ≤ Mo ≤ 270° SR > 0 HO

3.2.2. Vessels not in sight without radar

Identified in Rule 19 e), hearing plays the key role in this situation, although it is to be used in everyday
look-out duties according to COLREG 72, Rule 5.
• The vessel which hears apparently forward of her beam the fog signal has a Give-way status.

• The vessel which hears apparently NOT forward of her beam the fog signal has a Stand-on status.

The short range of audibility of sound signals, according to Rule 35 and Annex III of COLREG 72 (Table 6)
becomes another problem in this scenario.

Table 6. Sound signal: range of audibility.

LENGTH OF VESSEL (METERS) AUDIBILITY RANGE (NAUTICAL MILES)

200 or more 2

75 but less than 200 1.5

20 but less than 75 1

Less than 20 0.5

4. Other COLREG 72 criteria

As previously mentioned, COLREG 72 includes two other criteria in addition to visibility: Kind of Propulsion
and Restrictions.

4.1. Kind of propulsion

The means of propulsion recognized by COLREG 72 are shown in Table 7.

Table 7. Means of propulsion recognized by COLREG.

PROPULSION COLREG RULES Passage
preference

Power-driven vessel Rules 3, 10, 14, 15, 17, 18, 19, 23, 24, 27, 28, 34, 35. Annex I Rules 2 and 4; No
Annex III Rule 2 b)

Oars Rule 25 (Possible)

Sail Rules 3, 9, 10, 11, 12, 18, 25 and 35. Annex I Rule 10 Yes

Mix Sail and Power-driven Rule 25 (Possible)

Seaplanes Rules 3 and 31 No

Non-displacement craft Rules 3 and 23 No

Wing-In-Ground (WIG) Rules 3, 18, 23 and 31 No

craft

The type of propulsion could imply some passage preference, hence the importance of its identification.

This classification from COLREG 72 does not correspond to that from AIS, issue that could be avoided by standardizing
classifications.

4.2. Restrictions
COLREG identifies the limitations or restrictions on the normal functioning of vessels. Two groups of
restrictions can be differentiated. The first group, shown in Table 8, has to do with characteristics or activity
of the vessel. AIS dynamic messages include a navigational status flag related to these restrictions.

Table 8. Restrictions related to characteristics or vessel activities.

RESTRICTIONS COLREG RULES AIS NAVIGATIONAL STATUS

Vessel not under command Rules 3, 18, 27 and 35 Not under command

Vessel restricted in her ability to Rules 3, 10, 18, 27 and 35 Restricted in ability to
manoeuvre manoeuvre

Vessel engaged in mine clearance Rule 27

operations

Vessel engaged in fishing Rules 1, 3, 9, 10, 18, 26, 35; Annex I Rules 2 and 4; Engaged in fishing
Annex II Rule 3

Sailing vessel Rules 3, 9, 10, 12, 18, 25, 35; Annex I Rule 10 Underway by sail

Vessel constrained by her draught Rules 3, 18, 28 and 35 Constrained by draught

Anchored vessel or anchor Rules 3, 9, 10, 26, 27, 29, 30, 35; Annex I Rules 2 and 9 At anchor

Vessel aground Rules 3, 30, 35 Aground

RESTRICTIONS COLREG RULES AIS NAVIGATIONAL STATUS

Vessel made fast to the shore Rule 3 Moored

For the above restrictions there may be a number of acoustic signals, lights and/or markings that allow
visual or acoustic identification, within their range.

The second group of restrictions is based on geographical considerations (Table 9).

Table 9. Restrictions based on geographical considerations.

RESTRICTIONS COLREG RULES

Narrow channels (NC) Rule 9

Traffic separation schemes (TSS) Rule 10

Inshore traffic zone (ITZ) Rule 10

Traffic Lane (TL) Rule 10

There are neither marks nor lights that can be displayed by a vessel to indicate that she is in a restriction
based on geographical considerations.

Every vessel must obtain this information through the combined use of navigational aids: bouys, sound
signals, RADAR or AIS with nautical charts or ECDIS.

Ignorance of this situation can generate problems when two vessels encounter with different geographical
restrictions each other (Ever Smart-Alexandra1, 2015).

There is no definition at the level of international law in the case of the NC, as there is in the other three
cases: TSS, ITZ and TL, which are elements of Ship Routeing Systems (IMO, 1986b). In these areas, actions
may differ greatly from the general provisions (COLREG 72 Rules 8, 9 and 10).

5. Discussion
From reading the previous sections, it might seem that the COLREG analysis carried out is not important
enough to be the subject of a publication of this kind. However, an analysis of the articles on the subject in
high-impact journals is sufficient to realise that fundamental things that should be common for all those
working in this field do not seem to be so. We will show, with examples taken from influential or recent
publications on collision avoidance, some inaccuracies in the application of COLREG, which will adversely
affect the implementations derived from them.

5.1. Plotting the target SHIP (TS)

COLREG requirement for plotting the TS seems to be an issue that many authors struggle with. Usually, on
board, this is done with a reference system relative to the Own Ship (OS). In this way, the information
necessary for the classification of the situation is obtained by establishing a prediction of minimum passing
distance CPA or time of maximum approach. And so (CPA) it appears on ships' navigational equipment. The
use of the expression Distance at/to Closest Position of Approach, which is used in academic works, is
curious, as it is understood as a redundancy in the maritime professional field.

Plotting provides further information: the time at which that maximum approach will occur (TCPA), the
relative and the true movement of the TS.

Once all this information has been analyzed, the encounter can be assessed from the perspective of COLREG
Rule 8 (d): “Action taken to avoid collision with another vessel shall be such as to result in passing at a safe
distance … ".

It is important to know the potential of plotting as it covers all the information needs for collision
avoidance. There are statements that may be surprising, such as the following:
The model proposed in this paper incorporates some novelties regarding how an encounter between two
ships on a coastal restricted and open sea area is considered. Unlike most existing approaches, it combines
three variables relevant to collision risk, namely, the distance between encountering ships; the relative
speed of the ships; and the difference between the headings of the ships, referred to as the phase (Zhang et
al., 2015, p.67).

What it offers is exactly what plotting, as already described in Brown (1971), gives us.

The mentioned work from Zhang W. et al. also uses some measures well known in the maritime world and
whose limitations must be understood:
First, in certain head-encounters, if the other ship crosses ahead of own ship with a small value of the
closest point of approach (CPA) but sufficiently wide bow cross-range (BCR), the small CPA would not imply
an unsafe encounter. In such cases, the BCR is informative and can be used to make inferences, where the
CPA could have led to misclassification of a safe encounter as unsafe. (p.62)

One of the dangers that every seafarer has to be aware of is the false sense of reassurance that a safe BCR
can give over an unsafe CPA. There is no such thing as a safe encounter with an unsafe CPA.

Suppose different relative bearing for the same BCR (Fig. 2a). It can be seen that the CPA is always lower
than the BCR. The closer the vessel is to the bow (lower relative bearing), the lower the CPA. Similarly, for
the same CPA (Fig. 2b), the further the vessel is forward or in opposition, the higher the BCR will be.

Download : Download high-res image (243KB)

Download : Download full-size image
Fig. 2. CPAs and BCRs.

So, the BCR cannot be a reliable sign of a safe encounter.

In a recent publication (Zhang et al., 2021), the case shown in Fig. 3 does not comply with the behaviour
indicated in COLREG. Authors indicate that the OS needs to manoeuvre to move away from the TS.

Download : Download high-res image (677KB)

Download : Download full-size image

Fig. 3. Excerpt from Zhang X et al. (2021).

Analysing the proposed movements (Fig. 4) we obtain the passing distances shown in Fig. 5, in true (left)
and relative (right) movement. It can be seen how, without any manoeuvre, OS would not cross ahead TS as
advised by COLREG and would pass at a distance of 1.4 nm, which can be considered safe, so OS should not
manoeuvre.
Download : Download high-res image (203KB)
Download : Download full-size image

Fig. 4. Ship plotting from Fig. 3 data.

Download : Download high-res image (158KB)

Download : Download full-size image

Fig. 5. Passing distance.

It is also surprising that after the manoeuvre, which is estimated to be necessary, the passing distance is
about 0.30 nm.

5.2. Definition of encounters

Having defined an encounter as unsafe, whether collision or close-quarters, the next step is to analyse who
must initiate an evasive manoeuvre and how. It should be remembered that all vessels are obliged to
manoeuvre to avoid a collision, even if they have an initial situation of preference. COLREG provides answers
to these two questions. It is therefore very important not to simplify its content.

As already mentioned, literature mostly reflects the analysis of vessel in sight encounters, with head-on,
crossing and overtaking situations. It is also common the definition of fixed sectors, the application of which
any OONW knows cannot be generalised.

Sectors are never exclusive. As an example, possible encounters with relative bearings within the supposed
head-on sector on both sides of the vessel are given. Fig. 6 shows the three possible vessel in sight situations
with the same Relative Course. β, aspect, indicates the angle between the heading of the target vessel and
her relative bearing (βO = Overtaking aspect; βC = Crossing aspect; βH = Head-on aspect).
Download : Download high-res image (472KB)
Download : Download full-size image

Fig. 6. Vessels in sight. Multiple situations for the same relative bearing.

The concept of closed encounter sectors is striking and gives rise to considerable errors. The following
shows that the methodology presented in Du et al. (2021), adapted from Goerlandt et al. (2015) and Tam and
Bucknall (2010), two of the most cited publications in the collision avoidance literature (Gil et al., 2020), like
its predecessors is not effective in identifying potential collision situations and the corresponding
manoeuvring.

The ship COLREGs identity presented in Du et al. (2021) is determined according to the relative bearing (RB)
and relative heading (RH) (Fig. 7).

Download : Download high-res image (311KB)

Download : Download full-size image

Fig. 7. Regions and encounter array (Du et al., 2021).

There are some inaccuracies in the array of encounter identifications of OS. For example, a TS observed in
sector RB8, i.e. in [337.5°, 360°], with a relative Heading in sector RH2, [022.5°, 090°]. Although not
indicated, it is assumed that the boundaries are included in all sectors.

As can be seen in the table included in Fig. 7, authors state that any encounter combining RB8 and RH2 is a
CR/SO, i.e. the OS should maintain course and speed, at least initially, as it is considered a Stand On vessel in
a Crossing encounter scenario.

Fig. 8 shows a case included in these sectors, TS with RB = 348° (in the Fig. 012° = 360°–348°) and RH = 055.5°.
For relative bearings in RB8, RH = 055.5° would be the limit for a CR/SO encounter scenario. For lower RH, in
the range [022.5°, 055.5°[, the OS is a Give Way vessel in an Overtaking scenario (OT/GW), so the OS should
mano euvre.
Download : Download high-res image (406KB)
Download : Download full-size image

Fig. 8. TS in RB8, RH2.

Fig. 9 illustrates this affirmation, with an example without changing the Relative Bearing = 348° and with
RH = 040°.
Download : Download high-res image (288KB)
Download : Download full-size image

Fig. 9. TS with RB = 348°, RH = 040°.

It is an overtaking encounter, as OS visual enters through the sternlight sector of TS.

Also very surprising is the classification of encounters made for the intersection of sector RB1 with sectors
RH4 and RH5 (Fig. 7): (RB1, RH4) = CR/S0; (RB1, RH5) = HO/GW.

According to this classification, assuming a RB = 001° and RH = 179°, it is a CR/SO. However, with the same RB
but with a RH = 181° it is a HO/GW.

Limit values for the sectors RB1 and RB8 (π/8, 11π/8) are larger than those recommended, following the
indication in Tam and Bucknall (2010) to avoid changes from HO to CR with small variations in OS heading.
But, as shown in the previous paragraph, a change of 2° in TS heading (RH from 181° to 179°) would have the
same effect of switching from HO/GW to CR/SO.

As mentioned, similar problems with fixed encounter sectors can be found in Tam and Bucknall (2010) (Fig.
10).
Download : Download high-res image (279KB)
Download : Download full-size image

Fig. 10. Regions used to categorise position and heading of the obstacle (Tam and Bucknall, 2010).

From this article: “For example, if the obstacle is located in the region R2, and the heading of the obstacle is
in the zone TSR1, the resulting encounter type for the obstacle will be an overtaking (OT) encounter”
(p.260).

According to Fig. 10, R2 angles are in [22.5°, 90°] and TSR1 angles in [292.5°, 67.5°]. Fig. 11 shows the region
R2 including 3 examples of relative bearings with respect to the heading of the OS and the TSR1 zones for
the TS. For the encounter to be dangerous, the relative course must generate a situation of collision or close-
quarters. We have taken the collision option as the most obvious. In that case, relative courses are the
opposite of relative bearings. The range of possible TS headings to generate this dangerous situation is much
more limited than the TSR1 indicated by the authors. In a close-quarters situation it would not be so
restrictive, but it would not reach the proposed values either.

Download : Download high-res image (254KB)

Download : Download full-size image

Fig. 11. Examples with TS in R2, TSR1.

Furthermore, not all combinations of TSR1 with R2 generate an overtaking situation. As an example, Fig. 12
shows that for a relative bearing of 57.5° (relative course 237.5°) and TS heading in [292.5°, 350°] there is a
Crossing (CR) situation, not a OT.
Download : Download high-res image (178KB)
Download : Download full-size image

Fig. 12. Crossing encounter with TS in R2, TSR1.

On the other hand, Li et al. (2021), Rong et al. (2021) and Guo et al. (2020) coincide on the identification of 4
regions associated with the Head-on, Overtaking and Crossing situations (2) shown in Fig. 13.

Download : Download high-res image (206KB)

Download : Download full-size image

Fig. 13. Encounter regions (Rong et al., 2021).

As a counterexample, Fig. 14 shows a Crossing situation within the region associated with Head-on in the
aforementioned articles.
Download : Download high-res image (246KB)
Download : Download full-size image

Fig. 14. Crossing situation with relative bearing = 358°.

Guo et al. (2020) take into account encounters between vessels not in sight, although they interpret rule 19
in a way that is far from the COLREG statement, as shown in Table 10:

Table 10. Rule 19 and Guo et al. (2020) interpretation.

COLREG RULE 19 ARTICLE REMARKS

..so far as possible the following shall be In addition, in situation where the visibility of the sea COLREG say what to
avoided: environment is restricted, there will be no responsibility avoid, not what to
separation between the stand-on vessel and the give-way do.
vessel.
The COLREGS made the following rules for the situation of the
ship at this time:

(i) an alteration of course to port for a For coming ships in the range of (0°, 90°) and (270°, 360°), the It would be possible
vessel forward of the beam, other than self-ship will turn to the right. to change speed.
for a vessel being overtaken; It does not cover the
overtaking situation.

(ii) an alteration of course towards a Self-ship takes a turn towards other ships for coming ships in It is just the opposite.
vessel abeam or abaft the beam the range of (90°, 180°) and (180°, 270°).

Fortunately, they state that: “We mainly study the situation of good visibility at sea in this article” (p.13).

Another feature found in the literature is the variety of criteria for defining the Head-on sector (Table 11),
the strangest of which is found in a paper on encounter analysis sponsored by the IALA (2018) (Fig. 15).
Table 11. Head-on sectors in literature.

HEAD-ON SITUATION

M° (Port and Stbd) AUTHORS

5° (Cockcroft, 1982; Du et al., 2020; Tsou et al., 2010)

5.7° He et al. (2017)

10° Montewka et al. (2011)

15° Szlapczynski and Krata (2018)

22.5° Tam and Bucknall (2010)

30° IALA (2018)

Download : Download high-res image (233KB)

Download : Download full-size image

Fig. 15. Definition of the three encounter types (IALA, 2018).

Not a few studies adopt a Head On sector of (-π/8, +-π/8). The origin may lie in the work of Tam and Bucknall
(2010):
… there is no explicit guideline in Rule 14 of the COLREGs that defines a head-on encounter, except when
discussing the visibility of the masthead light and side-lights; sidelight visibility is defined in Annex I 9(a) of
the COLREGs to be small (1–3°). In this study, instead of the recommended values, the HO1 and HO2 values
were increased (to angles of π/8 radians), and they will be discussed later. (p.259)

We will see that these increased values for HO1 and HO2 will not allow compliance with COLREG Rule 14
(b):
Such a situation shall be deemed to exist when a vessel sees the other ahead or nearly ahead and by night
she could see the masthead lights of the other in a line or nearly in a line and/or both sidelights and by day
she observes the corresponding aspect of the other vessel.

It is important to highlight the fact that COLREG uses the OONW viewpoint, not a bird's-eye view. The visual
condition in a Head-on situation is established in Rule 14 (b) and the light horizontal sectors in Annex I.9.
Simulating the appearance would look as shown in Fig. 16: by night, masthead lights of the other in a line or
nearly in a line (1) and/or both sidelights (2). By day, the corresponding aspect of the other vessel (3).

Download : Download high-res image (291KB)

Download : Download full-size image

Fig. 16. Head-on situation. Rule 14 (b).

Fig. 17 shows the simulation of an encounter with relative bearing = 20° (i.e. 2.5° less than the sector
proposed by the authors and therefore within it).

Download : Download high-res image (208KB)

Download : Download full-size image

Fig. 17. Encounter with relative bearing = 20°.

According to the visual definition in COLREG Rule 14 (b), for a collision situation to occur in a Head-on
encounter, i.e. for the relative motion to give a CPA = 0′, the speed of the TS would have to be ∞, because for
both ships to see the lights perfectly aligned, a Heading/Course parallel to the relative bearing would be
required. Even allowing for the margin of error of the visible arcs of the lights and a moderate ship yaw, TS
Heading should be almost parallel to the relative bearing.

As the authors have not mentioned possible disturbance forces (wind, currents, waves) that would force us
to distinguish between Course and Heading, it can be assumed that the two coincide.
As an example, Fig. 18 shows that a TS speed approximately 6 times the OS speed would be required to
obtain that situation with small difference between relative bearing and target heading slopes. This mean
that for usual OS speeds between 12 and 15 knots, TS speed would be between 72 and 90 knots.

Download : Download high-res image (211KB)

Download : Download full-size image

Fig. 18. Head-on collision situation with relative bearing = 20°.

This is the first of the problems with setting wide Head-on angles on each side of the OS.

Also, to qualify the encounter as Head-on it has to be taken into account that both vessels must see the
same scenario.

From the OS, ignoring the magnitude of the speeds required, it could be accepted the TS is sighted according
to Rule 14(b), as shown in Fig. 19 (1).
Download : Download high-res image (245KB)
Download : Download full-size image

Fig. 19. Views from OS and TS

But it is clear that from the TS the view is different (Fig. 19 (2)).

Therefore, applying Rule 14 (b), the OS might qualify the encounter as Head-on but the TS would qualify it
as Crossing.

To conclude this section on the study of sectors, mention should be made of the recent article by Silveira et
al. (2021) (Fig. 20). In it, authors refer to the criterion of 5 experts that “decided to define six sectors with six
different risk levels” (p.6):

Download : Download high-res image (271KB)

Download : Download full-size image

Fig. 20. Sectors defined in Silveira et al. (2021).

They point out that: “Regarding the risk of collision, the experts’ judgement on the order of the sectors, from
higher risk to lower risk, is C, E, F, B, A, D”.

It is striking, particularly with this assessment of sector C as the most dangerous, that no account has been
taken of the fact that, within this sector, relative bearings in ]112.5°, 135°] correspond to an overtaking
encounter, while relative bearings in [78.5°, 112.5°] correspond to a crossing encounter on the starboard
side. In other words, two totally different types of encounters within the same sector.

If overtaking, OS would have to wait for TS to manoeuvre, in order to follow Rule 17, which can be a
distressing situation. If crossing on the starboard side, OS has to take the initiative, so the perception of risk
should be really different.

It seems that the “Duty to take action” criterion mentioned in the paper has not been taken into account for
the risk assessment.

6. Conclusions
COLREG 72 attempts to establish safe coordination in the movement of two vessels in a given close-quarters
or collision encounter. With the aim of being a universal and simplified guideline, it establishes a series of
rules that respond to the technical means of the time. It always seeks the essential character, avoiding the
superfluous, and includes a number of rules establishing more specific communication protocols: Rule 9 d)
e) and f); Rule 34, etc.

Any difference in the interpretation and application of COLREG 72 leads to an increased risk of collision in a
ship-to-ship encounter. New technological developments offer new forms of communication that can be
harnessed to enable a better understanding of the ships' intentions within the rules to be complied with. A
handshaking protocol for dialogues and agreements between the OONWs of both ships about the
rules/manoeuvres to apply can also be added, as in Argüelles et al. (2019; 2021), using the same COLREG
philosophy, short and unambiguous messages.

At present, DSS/CASs should be adapted to the requirements of COLREG 72 as a fundamental part of their
designs. It is necessary to apply the whole COLREG 72 and not just a part of it, nor to develop schemes based
on COLREG 60.

In this paper, basic criteria to consider when making decisions in vessel encounters in accordance with
COLREG have been identified and classified.

The results of the analysis of COLREG 72 call for immediate action to unify criteria for the calculations of
DSS/CAS algorithms, which the IMO should address as soon as possible:
- Disparity in head-on sectors,

- uncertainty in the interpretation of overtaking in not in sight vessels,

- lack of standardization of the classification of types of propulsion, and restrictions for all navigation aids,

- identification of special COLREG areas,

- ambiguity in the application of Rule 17,

are subjects that need urgent attention.

An analysis of how maritime traffic develops will be necessary. Many paths can be chosen. In any way IMO
should take homogeneous criteria for the regulation of navigable areas in order to reduce the possibility of
problematic encounters, especially with the emergence of MASS.

It is essential to add some inter-ship communications for data sharing in order to achieve safer encounters.
Vessels engaged in an encounter need to be aware of the fact that they are part of the same scenario and
situation, which might develop into a critical situation. Having the possibility of sharing and contrasting this
information helps making decisions in a consistent way.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that
could have appeared to influence the work reported in this paper.

Recommended articles

Data availability
The authors are unable or have chosen not to specify which data has been used.

References
ALLIANZ, 2020 ALLIANZ
Safety and shipping review 2020
[WWW Document]. URL
https://www.agcs.allianz.com/content/dam/onemarketing/agcs/agcs/reports/AGCS-Safety-Shipping-Review-
2020.pdf (2020)
9.15.21
Google Scholar

Antão and Guedes Soares, 2008 P. Antão, C. Guedes Soares

Causal factors in accidents of high speed craft and conventional ocean going vessels
Reliab. Eng. Syst. Saf., 93 (2008), pp. 1292-1304, 10.1016/j.ress.2007.07.010
View PDF View article View in Scopus Google Scholar

Belcher, 2002 P. Belcher

A sociological interpretation of the COLREGS
J. Navig., 55 (2002), pp. 213-224, 10.1017/S0373463302001686
View in Scopus Google Scholar

Bole et al., 2014 A. Bole, A. Wall, A. Norris

Radar and ARPA Manual Radar and Target Tracking for Professional Mariners, Yachtsmen
and Users of Marine Radar
 (third ed.), Butterworth-Heinemann (2014)
Google Scholar
Bowditch, 2019 N. Bowditch
American Practical Navigator. An Epitome of Navigation, ume I, National geospatial-intelligence agency (2019)
Google Scholar

Brown, 1971 E. Brown

Radar Navigation Manual (Pub 1310)
U.S. Naval Oceanographic Office (1971)
Google Scholar

Canal, 2015 Suez Canal

Rules of navigation
[WWW Document]. URL
https://www.suezcanal.gov.eg/English/Navigation/Pages/RulesOfNavigation.aspx (2015)
9.14.21
Google Scholar

Cockcroft, 1982 A.N. Cockcroft

The circumstances of sea collisions
J. Navig., 35 (1982), pp. 100-112
View in Scopus Google Scholar

Cockcroft and Lameijer, 2011 A.N. Cockcroft, J. Lameijer

A Guide to Collision Avoidance Rules
Butterworth-Heinemann (2011)
Google Scholar

Crosbie, 2009 J.W. Crosbie

Revisiting the lessons of the early steering and sailing rules for an e-navigation age
J. Navig., 62 (2009), pp. 109-117
View in Scopus Google Scholar

Du et al., 2020 L. Du, F. Goerlandt, O. Valdez Banda, Y. Huang, Y. Wen, P. Kujala

Improving stand-on ship's situational awareness by estimating the intention of the give-
way ship
Ocean. Eng. (2020), 10.1016/j.oceaneng.2020.107110
201
Google Scholar

Du et al., 2021 L. Du, O. Valdez Banda, Y. Huang, F. Goerlandt, P. Kujala, W. Zang

An empirical ship domain based on evasive maneuver and perceived collision risk
Reliab. Eng. Syst. Saf., 213 (2021), 10.1016/j.ress.2021.107752
Google Scholar
EMSA, 2020 EMSA
Annual overview of marine casualties and incidents 2020
[WWW Document]. URL
http://www.emsa.europa.eu/publications/reports/item/4266-annual-overview-of-marine-casualties-and-
incidents-2020.html (2020)
9.14.21
Google Scholar

Fujiwara et al., 2017 S. Fujiwara, S. Fujimoto, M. Fuchi, T. KOnishi

Gap between detailed information by navigational equipment and COLREGs rule 19
Fujiwara, Saeko & Fujimoto, Shoji & Fuchi, M. & KOnishi, Tsukasa
Trans. Navig., 2 (2017), pp. 25-33
View in Scopus Google Scholar

Gil et al., 2020 M. Gil, K. Wróbel, J. Montewka, F. Goerlandt

A bibliometric analysis and systematic review of shipboard Decision Support Systems for
accident prevention
Saf. Sci., 128 (2020), 10.1016/j.ssci.2020.104717
Google Scholar

Goerlandt and Kujala, 2011 F. Goerlandt, P. Kujala

Traffic simulation based ship collision probability modeling
Reliab. Eng. Syst. Saf., 96 (2011), pp. 91-107, 10.1016/j.ress.2010.09.003
View PDF View article View in Scopus Google Scholar

Goerlandt et al., 2015 F. Goerlandt, J. Montewka, V. Kuzmin, P. Kujala

A risk-informed ship collision alert system: framework and application
Saf. Sci., 77 (2015), pp. 182-204, 10.1016/j.ssci.2015.03.015
View PDF View article View in Scopus Google Scholar

Guo et al., 2020 S. Guo, X. Zhang, Y. Zheng, Y. Du

An autonomous path planning model for unmanned ships based on deep reinforcement
learning
Sensors, 20 (2020), 10.3390/s20020426
Google Scholar

He et al., 2017 Y. He, Y. Jin, L. Huang, Y. Xiong, P. Chen, J. Mou

Quantitative analysis of COLREG rules and seamanship for autonomous collision avoidance
at open sea
Ocean. Eng., 140 (2017), pp. 281-291, 10.1016/j.oceaneng.2017.05.029
View PDF View article View in Scopus Google Scholar

Hilgert and Baldauf, 1997 H. Hilgert, M. Baldauf

A common risk model for the assessment of encounter situations on board ships
 Ger. J. Hydrogr., 49 (1997), pp. 531-542
Google Scholar
IALA, 2018 IALA
Report 18591.620/TECH_DOC/2. Encounter Model
(2018)
Google Scholar

IMO, 1986a IMO

Performance standards for automatic radar plotting aids (ARPAs)
Resolut. A, 823 (19) (1986)
Google Scholar

IMO, 1986b IMO

General provisions on ships' routeing
Resolut. A, 572 (14) (1986)
Google Scholar

IMO, 2001 IMO

Standard marine communication Phrases
Resolut. A, 918 (22) (2001)
Google Scholar

IMO, 2004 IMO

Performance standards for the presentation of navigation-related information on
shipborne navigational displays
Resolut. MSC, 191 (79) (2004)
Google Scholar

IMO, 2015 IMO

Resolution A.1106. Revised Guidelines for the Onboard Operational Use of Shipborne
Automatic Identification Systems (AIS)
(2015)
Google Scholar

Kemp, 2009 J. Kemp

Behaviour patterns in crossing situations
J. Navig., 62 (2009), pp. 443-453, 10.1017/S0373463309005293
View in Scopus Google Scholar

Li et al., 2021 L. Li, D. Wu, Y. Huang, Z.-M. Yuan

A path planning strategy unified with a COLREGS collision avoidance function based on
deep reinforcement learning and artificial potential field
Appl. Ocean Res., 113 (2021), 10.1016/j.apor.2021.102759
Google Scholar
Lopez-Santander and Lawry, 2017 A. Lopez-Santander, J. Lawry
An ordinal model of risk based on mariner's judgement
J. Navig., 70 (2017), pp. 309-324, 10.1017/S0373463316000576
View in Scopus Google Scholar

Mohovic et al., 2016 D. Mohovic, R. Mohovic, M. Baric

Deficiencies in learning COLREGs and new Teaching methodology for nautical engineering
students and seafarers in lifelong learning programs
J. Navig., 69 (2016), pp. 765-776, 10.1017/S037346331500096X
View in Scopus Google Scholar

Montewka et al., 2010 J. Montewka, T. Hinz, P. Kujala, J. Matusiak

Probability modelling of vessel collisions
Reliab. Eng. Syst. Saf., 95 (2010), pp. 573-589, 10.1016/j.ress.2010.01.009
View PDF View article View in Scopus Google Scholar

Montewka et al., 2011 J. Montewka, P. Krata, F. Goerlandt, A. Mazaheri, P. Kujala

Marine traffic risk modelling - an innovative approach and a case study
Proc. Inst. Mech. Eng. Part O J. Risk Reliab. (2011), pp. 225-307, 10.1177/1748006X11399988
Google Scholar

Panama Canal, 2019 Panama Canal

Regulation on navigation in Panama canal waters
Marit. Regul. (2019)
[WWW Document]. URL
https://www.pancanal.com/eng/maritime/regulations/index.html
9.14.21
Google Scholar

Perera et al., 2009 L.P. Perera, J.P. Carvalho, C.G. Soares

Autonomous guidance and navigation based on the COLREGs rules and regulations of
collision avoidance
Proceedings of the International Workshop “Advanced Ship Design for Pollution Prevention (2009), pp. 205-216
Google Scholar

Perera et al., 2010 L.P. Perera, J.P. Carvalho, C. Guedes Soares

Fuzzy-logic based parallel collisions avoidance decision formulation for an Ocean
Navigational System
IFAC Proceedings Volumes (2010), pp. 260-265, 10.3182/20100915-3-DE-3008.00044
View PDF View article View in Scopus Google Scholar

Perera et al., 2011 L.P. Perera, J.P. Carvalho, C. Guedes Soares

Fuzzy logic based decision making system for collision avoidance of ocean navigation
under critical collision conditions
 J. Mar. Sci. Technol., 16 (2011), pp. 84-99, 10.1007/s00773-010-0106-x
View in Scopus Google Scholar
Rong et al., 2021 H. Rong, A.P. Teixeira, C. Guedes Soares
Spatial correlation analysis of near ship collision hotspots with local maritime traffic
characteristics
Reliab. Eng. Syst. Saf., 209 (2021), 10.1016/j.ress.2021.107463
Google Scholar

Salinas, 2006 C.F. Salinas

Restricted visibility: in search of a solution
J. Navig., 59 (2006), pp. 352-363
Google Scholar

Silveira et al., 2021 P. Silveira, A.P. Teixeira, J.R. Figueira, C. Guedes Soares
A multicriteria outranking approach for ship collision risk assessment
Reliab. Eng. Syst. Saf., 214 (2021), 10.1016/j.ress.2021.107789
Google Scholar

Smart-Alexandra1, 2015 Ever Smart-Alexandra1

Report on the investigation of the collision between the container ship Ever Smart and the
oil tanker Alexandra 1
[WWW Document]. URL
https://assets.publishing.service.gov.uk/media/5665aff8e5274a0367000010/MAIBInvReport-28_2015.pdf
(2015)
9.14.21
Google Scholar

Smeaton and Coenen, 1990 G.P. Smeaton, F. Coenen

Developing an intelligent marine navigation system
Comput. Control Eng. J., 1 (1990), pp. 95-103
CrossRef View in Scopus Google Scholar

Szlapczynski and Krata, 2018 R. Szlapczynski, P. Krata

Determining and visualizing safe motion parameters of a ship navigating in severe
weather conditions
Ocean. Eng., 158 (2018), pp. 263-274, 10.1016/j.oceaneng.2018.03.092
View PDF View article View in Scopus Google Scholar

Szlapczynski et al., 2018 R. Szlapczynski, P. Krata, J. Szlapczynska

Ship domain applied to determining distances for collision avoidance manoeuvres in give-
way situations
Ocean. Eng., 165 (2018), pp. 43-54, 10.1016/j.oceaneng.2018.07.041
View PDF View article View in Scopus Google Scholar
Tam and Bucknall, 2010 C. Tam, R. Bucknall
Collision risk assessment for ships
J. Mar. Sci. Technol., 15 (2010), pp. 257-270, 10.1007/s00773-010-0089-7
View in Scopus Google Scholar

Tsai et al., 2017 C.-C. Tsai, J.-R. Chang, C.-L. Chen

Manoeuvrability-based critical time for preventing close-quarters situations
J. Mar. Sci. Technol., 25 (2017), pp. 249-258, 10.6119/JMST-016-1207-2
View in Scopus Google Scholar

Tsou et al., 2010 M.-C. Tsou, S.-L. Kao, C.-M. Su

Manoeuvrability-based critical time for preventing close-quarters situations
J. Navig., 63 (2010), pp. 167-182, 10.1017/S037346330999021X
View in Scopus Google Scholar

Wang et al., 2017 X. Wang, Z. Liu, Y. Cai

The ship maneuverability based collision avoidance dynamic support system in close-
quarters situation
Ocean. Eng., 146 (2017), pp. 486-497, 10.1016/j.oceaneng.2017.08.034
View PDF View article View in Scopus Google Scholar

Wróbel, 2021 K. Wróbel

Searching for the origins of the myth: 80% human error impact on maritime safety
Reliab. Eng. Syst. Saf., 216 (2021), 10.1016/j.ress.2021.107942
Google Scholar

Xu and Wang, 2014 Q. Xu, N. Wang

A survey on ship collision risk evaluation
Promet - Traffic & Transp., 26 (2014), pp. 475-486, 10.7307/ptt.v26i6.1386
View in Scopus Google Scholar

Zhang et al., 2015 W. Zhang, F. Goerlandt, J. Montewka, P. Kujala

A method for detecting possible near miss ship collisions from AIS data
Ocean. Eng., 107 (2015), pp. 60-69, 10.1016/j.oceaneng.2015.07.046
View PDF View article Google Scholar

Zhang et al., 2018 G. Zhang, V. Thai, K.F. Yuen, H.S. Loh, Q. Zhou
Addressing the epistemic uncertainty in maritime accidents modelling using Bayesian
network with interval probabilities
Saf. Sci., 102 (2018), pp. 211-225, 10.1016/j.ssci.2017.10.016
View PDF View article View in Scopus Google Scholar

Zhang et al., 2021 X. Zhang, C. Wang, K.T. Chui, R.W. Liu

 A real-time collision avoidance framework of MASS based on B-spline and optimal
decoupling control
Sensors, 21 (2021), 10.3390/s21144911
Google Scholar

Cited by (6)

Research on risk, safety, and reliability of autonomous ships: A bibliometric review

2023, Safety Science

Show abstract

Development of ship collision avoidance system and sea trial test for autonomous ship
2022, Ocean Engineering

Citation Excerpt :
…In addition, COLREGs are international rules designed with the navigator as the center of judgment to avoid collisions
between ships. Therefore, analysis of collision situations different from the navigator’s point of view should be avoided
(Maza and Argüelles, 2022). In this paper, the experiment was limited to situations encountered in head-on and crossing,
and the experiment was conducted to follow the rules of COLREGs.…

Show abstract

A Decision Support System Using Fuzzy Logic for Collision Avoidance in Multi-Vessel Situations
at Sea
2023, Journal of Marine Science and Engineering

Research on MASS Collision Avoidance in Complex Waters Based on Deep Reinforcement

Learning
2023, Journal of Marine Science and Engineering

Polygonal Type Fuzzy Ship Domain-Based Decision Support System for Collision Avoidance
Route Planning
2023, Journal of Eta Maritime Science

The Vagueness of COLREG versus Collision Avoidance Techniques—A Discussion on the Current
State and Future Challenges Concerning the Operation of Autonomous Ships
2022, Sustainability (Switzerland)

© 2022 The Author(s). Published by Elsevier Ltd.

All content on this site: Copyright © 2023 Elsevier B.V., its licensors, and contributors. All rights are reserved, including those for text and data mining, AI training, and similar
technologies. For all open access content, the Creative Commons licensing terms apply.