Conference PaperPDF Available

Situation Awareness of Autonomous Ship Navigation in a Mixed Environment Under Advanced Ship Predictor

Authors:

Abstract and Figures

Autonomous ship navigation in a mixed environment, where remote-controlled, autonomous and manned vessels are interacting, is considered. These vessels can have various encounter situations, therefore adequate knowledge on such situations should be acquired to take appropriate navigation actions. That has often been categorized as situations awareness in a mixed environment, where appropriate tools and techniques to improve the knowledge on ship encounter situations should be developed. Hence, possible ship collision and near-miss situations can be avoided by both humans as well as systems. The collision risk assessment has an important role in ship situations and that can eventually be used towards the respective collision avoidance actions. Ship collision avoidance actions are regulated by the International Regulations for Preventing Collisions at Sea 1972 (COLREGs) in open sea areas and additional local navigation rules and regulations can also enforce especially in confined waters and maritime traffic lanes in these vessels. It is expected that the COLREGs and other navigation rules and regulations will be interpreted by both humans as well as systems in future vessels and those interpretations will be executed in to collision avoidance actions. Therefore, adequate understanding on situation awareness should be achieved to overcome possible regulatory failure due to human and system decisions, i.e. avoid possible collision or near-miss situations in a mixed environment. This study focuses on identifying such challenges in future ship encounters with possible solutions to improve situation awareness in a mixed environment.
Content may be subject to copyright.
1 Copyright © 2019 by ASME
Proceedings of the 38th International Conference on Ocean, Offshore and Arctic Engineering
OMAE2019
June 9-14, 2019, Glasgow, Scotland, UK
OMAE2019-95571
SITUATION AWARENESS OF AUTONOMOUS SHIP NAVIGATION IN A MIXED
ENVIRONMENT UNDER ADVANCED SHIP PREDICTOR
Lokukaluge P. Perera and Brian Murray
UiT The Arctic University of Norway, Tromso, Norway
{Prasad.Perera, Brian.J.Murray}@uit.no
ABSTRACT
Autonomous ship navigation in a mixed environment, where
remote-controlled, autonomous and manned vessels are
interacting, is considered. Since these vessels can have various
encounter situations, adequate knowledge on such situations
should be acquired to take appropriate navigation actions. That
has often been categorized as situation awareness in a mixed
environment, where appropriate tools and techniques to extract
the respective knowledge on ship encounter situations should be
developed. The collision risk assessment procedure has an
important role in the same knowledge and that can eventually
be used towards the respective collision avoidance actions.
Hence, possible ship collision and near-miss situations can be
avoided by both humans as well as systems due to their actions.
Ship collision avoidance actions are regulated by the
International Regulations for Preventing Collisions at Sea 1972
(COLREGs) in open sea areas and additional local navigation
rules and regulations can also enforce especially in confined
waters and maritime traffic lanes. It is expected that the
COLREGs and other navigation rules and regulations will be
interpreted by both humans as well as systems in future vessels
and those interpretations will be executed as collision avoidance
actions by the respective vessels in a mixed environment.
Adequate understanding on situation awareness should be
achieved to overcome possible regulatory failure due to human
and system decisions in these situations. Hence, this study
focuses on identifying such challenges in future ship encounters
with possible solutions to improve situation awareness in a
mixed environment as the main contribution.
INTRODUCTION
Ship Navigation Safety
The accuracy of the collision risk assessment is an important
factor in evaluating situation awareness in ship encounter
situations [1]. Since collision avoidance actions of ship
navigators completely depend on the risk of possible collisions
or near-miss situations (i.e. within a considerable confidence
interval), that can play an important role in their decision-
making processes. However, the decision-making processes of
ocean-going vessels can further be complicated in the future
due to encounter situations of remote-controlled, autonomous
and manned ships. These have often been categorized as ship
encounters in a mixed environment. Appropriate tools and
techniques to extract the respective knowledge on ship
encounter situations in a mixed environment should be
developed, such that possible ship collision and near-miss
situations can be avoided. This approach should consist of
adequate measures to evaluate situation awareness including the
collision risk in vessel encounters to improve the respective
navigation safety. Hence, this study investigates such challenges
in future ship encounters along with the tools and techniques
that can improve situation awareness in a mixed environment as
the main contribution.
COLREGs & Collision Risk
Ship encounters that relate to possible near-miss and collision
situations are regulated by the International Regulations for
Preventing Collisions at Sea 1972 (COLREGs) [2] in open sea
areas. Furthermore, additional local navigation rules and
regulations can be enforced in ship navigation, especially in
confined waters and maritime traffic lanes. The COLREGs
classify the respective vessels in a close encounter situation, i.e.
as stand-on and give-way vessels, where the stand-on vessel
should hold its course and speed, with the highest priority for
navigation and the give-way vessel (i.e. with the less priority for
navigation) required to keep out of the way of the stand-on
vessel. One should note that in any ship encounter situation, if
there is no collision risk, then the vessels should continue their
course and speed conditions, irrespective of the crossing
2 Copyright © 2019 by ASME
direction accordance with the COLREGs. That can be further
illustrated as: if there are no collision risk, then there will not be
any give-way or stand-on vessels in ship encounter situations;
therefore none of the vessels should change their course and
speed conditions and/or the respective navigation trajectories.
This is also an important concept to consider, since the change
of course and speed conditions in one vessel can cause a
possible collision or near-miss situation and such actions can
confuse other vessels that are following the COLREGs rules
and regulations. On the other hand, if the collision risk can be
detected relatively far away from a ship encounter situation,
then both vessels can take appropriate actions, irrespective of
the COLREGs rules and regulations. That can be done by a
slight change of vessel course and/or speed to reduce the
respective collision risk and that eliminates the possibility of
any close ship encounter situations. Even though this is an
important and simple research concept to consider, i.e. the early
detection and elimination of close ship encounter situations with
a slight change of course or speed, it has been ignored by a
majority of collision avoidance algorithms and studies.
One should note that it is extremely difficult to create a
vessel collision situation in realistic ocean-going conditions, i.e.
due to the same reason of a slight change of course or speed
actions can completely eliminate the respective collision or
near-miss situation. Such course and speed changes can occur
not only due to navigator’s actions, but also to external
environmental conditions, i.e. wind and wave conditions. In
general, ship encounter situations can be quantified by the
respective collision risk. To quantify a ship encounter situation
as a possible collision or near-miss situation, the respective
collision risk level should hold continuously. In general,
increasing or decreasing collision risk conditions can eliminate
possible collision or near-miss situations. Therefore, this
situation can also be an extreme or optimal situation, i.e. to hold
a continuous collision risk level, due to the same reason of a
slight change of course or speed actions can eliminate the
respective collision risk, completely. Vessels can often be
influenced by various environmental factors i.e. wind and wave,
etc., therefore constant collision risk levels may not be possible
to occur as mentioned before. The collisions and near-misses in
ship navigation can be infrequent situations due to the same
reasons and a slight change of course and speed changes can be
used to avoid such situations.
Comprehensive course and speed control actions, i.e.
such as path planning type approaches, may therefore not be
required to avoid collision or near-miss situations. On the other
hand, such approaches may also violate the COLREGs. That
can be taken as an important concept to consider for the
respective collision avoidance approaches in ship encounter
situations. If the vessels have taken appropriate collision
avoidance actions when they are in a possible collision or near-
miss situation, that can mitigate the respective collision risk.
Since the collision risk has been eliminated from the ship
encounter situation, then the vessels should keep their course
and speed unchanged to eliminate any possible future collision
or near-miss situations.
On the other hand, any course and speed changes can
lead towards another collision or near-miss situation and that
can also be a violation of the COLREGs. Furthermore, if the
vessels are in a close encounter situation without the collision
risk or eliminated collision risk, then both vessels should
continue their course and speed conditions in accordance with
the COLREGs. Any course and speed change actions can be
violations of the COLREGs resulting in a possible collision and
near-miss situation. Therefore, the navigator actions in a close
ship encounter situation without an adequate collision risk
assessment procedure can result in possible collisions or near-
miss situations. However, this simple concept, i.e. an
appropriate collision risk assessment procedure to support
collision avoidance actions, has also been ignored by many
research studies, where rather complex algorithms based on
sophisticated optimality conditions are proposed. It is a
possibility that such optimization algorithms can often violate
the COLREGs. On the other hand, those approaches can be
neither realistic nor computationally effective enough to be
implemented on-board vessels.
When onboard systems are making a considerable
amount of navigation decisions in future vessels, i.e.
autonomous vessels, those algorithms can further complicate
ship encounters in a mixture environment. On the other hand,
non-collision risk situations can be transformed into possible
collision or near-miss situations due to the same reasons. As a
possible solution, the collision risk among vessels should be
monitored continuously for both long-range and close ship
encounters as a part of situation awareness in future ships.
Therefore, this study considers developing tools and techniques,
a so called solution framework, to detect and monitor the
respective collision risk among vessels in long-range (i.e. global
scale) and close (i.e. local scale) ship encounters. Furthermore,
the applicability and evaluation procedures of such a solution
framework in future ship navigation in a mixed environment are
also further discussed.
Ship Under-Actuation
It is well known that ocean-going ships are under-actuated
systems, because these vessels may not have the full
controllability of their motions. Ship under-actuation can further
complicate close encounter situations due to unexpected
motions, especially under rough weather conditions. Hence, a
navigator should always be on the bridge to evaluate/re-evaluate
the collision risk with respect to ship behavior, i.e. resulted due
his course and speed control actions. Ship under-actuation can
also complicate the navigator`s decision-making process in such
situations, where adequate understating of situation awareness
may not be able to achieve. While systems are making
navigation decisions in future vessels, situation awareness
should be a part of their system intelligence to overcome the
same challenges. The voyage in a future ship, i.e. an
3 Copyright © 2019 by ASME
Fig. 1. Ship encounter situation under the solution framework.
autonomous ship, may consist of both manned (i.e. remote
controlled) and unmanned voyage legs. When onboard systems
(i.e. unmanned voyage legs) are making navigation decisions in
autonomous vessels, ship under-actuation can further
complicate not only those system decisions but also their
interactions (i.e. the outcomes) with the decisions made by
manned vessels (i.e. human decisions). Since ship under-
actuation that can result unexpected and undesirable vessel
motions in such encounter situations, adequate measures to
understand and quantify the situation awareness should be
considered.
Since future vessels will have both autonomous and
remote-controlled capabilities, ship under-actuation should also
be further investigated under not only from the system
perspective but also from the regulatory perspective. One
should note that the complete controllability of ocean-going
ships is something impossible to achieve, therefore adequate
tools and techniques to predict such unexpected and undesirable
motions, i.e. due to ship under-actuation, can help future vessels
to prepare for such situations. The same will further improve
the safety in ship navigation as a part of situation awareness. It
is also believed that the proposed solution framework, i.e. to
detect and monitor the respective collision risk among vessels
in local and global scales, can provide an elegant solution to
ship under-actuation by predicting such unexpected and
undesirable ship motions. Hence, both humans and systems can
take appropriate navigation decisions to cope with such
motions, where the safety levels of the respective ship
encounter situations can be further improved.
Research Challenges
Even though these issues can be crucial to improve situation
awareness in ship encounters, they have been overlooked by the
research community, as mentioned before. Hence, this study
incorporates such issues into situation awareness (i.e. including
the collision risk assessment process) of ship encounters in a
mixed environment under the proposed solution framework. In
addition, how to integrate the complexities in ship behavior, i.e.
ship seakeeping and maneuvering, into the collision risk
assessment procedures have also been considered under the
proposed advanced ship predictor. The outcome of the proposed
framework can improve ship collision avoidance actions that
can be taken by humans as well as systems during vessel
encounter situations. Furthermore, the same outcome can also
be used to identify possible regulatory failures in ship
navigation, where appropriate regulatory modifications under
the same framework can be proposed. This can improve the
COLREGs and other local navigation rules and regulations,
when applied towards manned, remote-controlled and
autonomous vessels. Furthermore, the same solution framework
can contribute to knowledge creation and competence
development in the research area of the safety and risk
assessment in mixed environmental operations, where manned,
remote-controlled, and autonomous vessel encounters can
occur.
THE SOLUTION FRAMEWORK
Advanced Ship Predictor
The proposed solution framework to address the respective
research challenges is presented in Figure 1 with the special
focus on close ship encounter situations in a mixed
environment. This framework is developed under the research
project of UiT Autonomous Ship Program. That starts with
analyzing the COLREGs and local navigation rules and
regulations and understanding their contributions to avoid ship
collisions. Ship encounter situations in open sea areas and high
traffic shipping lanes can be a part of such a framework in
relation to the regulatory analysis of the COLREGs and local
navigational rules and regulations. The outcome of such a
regulatory analysis can be used towards the proposed situation
awareness module (SAM). The two main components of the
SAM considered (see Figure 2): the collision risk assessment
unit (CRAU) and advanced ship predictor unit (ASPU).
Advanced Ship Predictor
The ASPU is proposed as a solution to ship under-actuation,
where the positions and orientations ahead of time in multiple
vessels can be estimated. These vessel position and orientation
estimations can be done by collecting onboard sensors in a local
scale and AIS data in a global and then fusing this information
by considering appropriate ship maneuvering models [3] with
estimation algorithms [4]. Since future vessels will be
facilitated by various sensors, this approach can be seen as a
method of harvesting ship seakeeping and maneuvering
information from the respective data sets and that will also be a
part of situation awareness in ship navigation. One should note
that that unexpected vessel behavior due to ship under-actuation
can be captured by the ASPU by predicting possible future
4 Copyright © 2019 by ASME
vessel positions and orientations with adequate confidence
intervals. When future vessel positions and orientations can be
predicted for a ship encounter situation, the same information
can be used to predict possible ship collision and near-miss
situations under the CRAU. Therefore, the outcome of the
ASPU can facilitate the CRAU to enhance its risk estimation
capabilities under the proposed solution framework.
The CRAU consists of detecting a collision or near-
miss situation for a two vessel encounter and these two vessel
encounter situations can be extrapolated towards a multi-vessel
encounter situation. In general, one vessel’s relative navigation
trajectory with respective to another vessel can be considered to
determine the respective collision risk. That can also be
identified as the predicted relative trajectory of one vessel with
the respective other vessel. The violation of the ship domain of
one vessel with respect to another vessel’s relative navigation
trajectory can be categorized as a possible collision or near-
miss situation. That violation can be quantified to estimate the
respective collision risk in the ship encounter situation. Such an
approach has been implemented by the previous studies and that
concept has been adopted towards this solution framework.
Furthermore, the collision risk between two vessels should be
evaluated with respect to not only the time but also the distance
until a possible collision or near-miss situation [5]. The
combination of the time and distance until possible collision or
near-miss situation can further improve the existing collision
risk assessment procedures.
The respective collision avoidance actions in remote-
controlled, autonomous, and manned vessels can be supported
by the outcome of the CRAU, where adequate information on
future ship behaviour should be communicated (i.e. by the
ASPU). One should note that the information coming from the
CRAU should be shared among the vessels in the respective
encounter situation, i.e. by integrated bridge systems.
Therefore, the same information should be obtained by the
following three navigation groups to improve situation
awareness: on-board humans in manned vessels, shore-based
humans in remote-controlled vessels and on-board systems in
autonomous vessels. Each group can have their own
perspectives on the ship encounter situation, even though they
are receiving the same information. However, the
decision/action differences among remote-controlled,
autonomous, and manned vessels for each collision situation in
a mixed environment are yet to be investigated in the future.
The combination of the ASPU and CRAU, with the respective
collision avoidance actions, completes the SAM. This can be a
standard framework to overcome the respective challenges in
situation awareness of close ship encounter situations in a
mixed environment consisting an appropriate collision risk
assessment procedure, i.e. based on advanced ship predictor,
under the COLREGs and local navigation rules and regulations.
Experimental Evaluations
This framework, i.e. the SAM, should be evaluated with various
ship encounter situations of remote-controlled, autonomous and
manned vessels [6] under both bridge simulator conditions and
realistic ocean-going conditions. However, adequate research
infrastructure, including bridge simulators, ocean going vessels
and human resources are required to evaluate such a solution
framework. Modern bridge simulators are facilitated with
adequate features to simulate and study various ship encounter
situations and that can be an appropriate and less expensive
platform to achieve the expected outcomes. It is expected that
future ship navigation will also be done by shore-based control
centers, therefore the usage of bridge simulators can create an
initial step towards such an approach. That can also satisfy
future training requirements, where the required shore-based
navigators, i.e. in remote-control centers, for autonomous ships
should be trained. Furthermore, various ship types and
environmental conditions can be introduced in this environment
and the vessels can be controlled by both remote-controlled and
on-board modes under the supervision of experienced ship
navigators.
An example of such an experiment is described in this
section (see Figure 2 for the SAM in a bridge simulator). Bridge
simulators may have the facilities to create multi-vessel
encounter situations with several simulated vessels. E.g. Vessel
A, B, C and D, etc.. Each vessel can have its own navigation
and control unit/station. Ship navigators should be placed in
these in these navigation and control units to create remote-
controlled and manned vessel navigation situations. The
behaviour of an autonomous ship in the same vessel encounter
situation can be created by a system that consists of the Parallel
Decision-making Module (PDMM) and Sequential Action
Formulation Module (SAFM). The PDMM module creates
appropriate collision avoidance decisions, i.e. course and speed
changes, by considering the collision risk estimated from the
CRAU and the COLREGs and local navigation rules and
regulations. The SAFM arranges those collision avoidance
decisions into appropriate actions in a sequential format that
can be implemented in ship rudder and propeller control
systems with respect to an appropriate timeline. These modules
that have been developed and experimented in simulations and
limited model scale vessel experiments [7] should complement
to each other as a decisionaction execution model to facilitate
intelligent collision avoidance features into autonomous vessels,
while respecting the COLREGs and local navigation rules and
regulations.
One should note that various computational
intelligence approaches have also been used to transform the
COLREGs and local navigation rules and regulations into
system readable formats. The proposed modules are one
approach that have been implemented before, therefore the
same can be implemented as a decisionaction execution model
for autonomous ship navigation. Furthermore, additional
computational intelligence approaches to improve such a
5 Copyright © 2019 by ASME
decisionaction execution model should be investigated. These
approaches should support if-then-else type conditions, since
the COLREGs and local navigation rules and regulations are
often following a similar logic. The decisionaction execution
model (i.e. PDMM and SAFM) should be connected to the
bridge simulator during these experiment evaluations to imitate
a digital helmsman in autonomous ship navigation.
Fig. 2. The proposed situation awareness module (SAM) in a bridge simulator.
It is expected that the outcome of bridge simulator
experiments can consist of various ship collision and near-miss
situations, where possible regulatory failures can occur. The
regulatory failures that can be observed under the experimental
results should be further be investigated and the outcome can be
used to introduce required regulatory modification into the
COLREGs and other local navigation rules and regulations.
Furthermore, the same outcome can also be used to improve the
SAM by considering the respective lessons learned during these
experiments.
Ship Encounter Situations
The evaluation of the same framework under vessel encounters
in realistic ocean-going conditions can be an extremely difficult
task due to the complexities in ship navigation. Such
complexities of a two-vessel encounter situation are discussed
in this section and that may give one an overview of the
respective issues that should be investigated in the future. A
two-vessel close encounter situation is presented in Figure 3
with the positions of Pa (xa(t), ya(t)) and Pb (xb(t), yb(t)) for
vessel A and B, respectively. The course-speed vectors of both
vessels denoted by Va(t) and Vb(t) and the heading vectors, i.e.
surge velocity vectors, are denoted by ua(t) and ub(t). An
imaginary position Pab is also denoted in the same figure to
represent a possible collision or near-miss situation. As
discussed before, ocean-going vessels are considered as under-
actuated systems with respect to their ship seakeeping and
manoeuvring behaviour. Therefore, such vessels may not have
complete controllability in the sway direction, i.e. denoted by
va(t) and vb(t), with propeller-rudder control systems (i.e.
resulting in ship under-actuation behaviour). The ship course-
speed and heading vectors can have two separate directions due
to the same reason and that can complicate the controllability of
these vessels due to undesirable and unexpected vessel motions.
The external forces, i.e. hydrodynamic and wind forces
and moments, influence on the sway direction of vessels, can
complicate seakeeping and manoeuvring behaviour [8]. Even
though rudder and thruster systems are installed in these vessels
to overcome such under-actuation behaviour, i.e. by providing
enough force against the sway direction, those systems may not
produce adequate thrust to control vessels when ship speeds are
over 3-4 knots. In general, a majority of ship navigation
situations are beyond this speed range. These sway velocity
component effects are neglected by a majority of research
studies, i.e. by assuming fully actuated vessels. Therefore, that
may not represent realistic ship navigation situations.
Furthermore, the solution provided by such studies may not be
applicable for ship navigation situations with vessel under-
actuation. Hence, ship encounter situations can have such
complex navigational challenges with compared to other
transport systems and that have not been addressed by the
research community [9], adequately.
The same under-actuation behaviour can complicate
the encounter situations, especially with remote-controlled,
autonomous and manned vessels. Since ship on-board systems
in the future will make a considerable amount of navigation
decisions, i.e. based on machine learning and artificial
intelligence, ship under-actuated behaviour can introduce
additional challenges. Furthermore, the interactions between
system and human decisions in the same environment can
further complicate ship encounter situations. On the other hand,
the human and machine perceptions on the same collision
avoidance rules and regulations can vary, therefore the collision
avoidance actions can be different. However, that difference
can also increase the risk of ship collisions considerably and
result in ship collisions and near-misses including possible
regulatory failures. Such challenges in ship encounter situations
in a mixed environment have not been studied in realistic
ocean-going conditions previously. Therefore, this study
proposes to investigate such fundamental requirements and
develop a realistic collision risk assessment procedure in vessel
navigation, e.g. including ship under-actuation behaviour, as a
part of situation awareness.
ADVANCED SHIP PREDICTOR
Ship Prediction
Predicting ship behaviour [10], accurately ahead of time in a
close ship encounter situation, can support the decision maker,
i.e. human or system, to take appropriate collision avoidance
6 Copyright © 2019 by ASME
Fig. 3. Ship encounter situation.
actions, and can also be a solution to ship under-actuation
behaviour. Therefore, unexpected ship motions due to vessel
under-actuation can be observed as the collision risk and
adequate actions by both humans and systems can be taken to
prevent possible collision and near-miss situations. The
collision risk assessment procedure [11] should be facilitated by
an advanced ship predictor methodology, i.e. the ASPU, that
estimates the future positions and orientations of the vessels in
the respective encounter situations. Therefore, the ASPU can be
an integrated part of the collision risk assessment procedure to
support situation awareness.
The on-board sensors in ships can collect the
information on vessel translational and rotational motions.
Furthermore, AIS (i.e. Automatic identification system) data
[12] of ocean-going vessels can be used to collect the
information on ship navigation behaviour. Therefore, this study
proposes an advanced ship predictor methodology, i.e. the
ASPU, that uses on-board sensor and AIS data to extract the
required information and predict future ship behaviour. Such
ship behaviour prediction can be done in two scales: local scale
and global scale. These two scales in predicting ship behaviour
are further elaborated in the following sections.
Local Scale
Various kinematic, dynamic or combined models (both
kinematics and dynamics) have been considered for ship
manoeuvring. However, the dynamic models in ship
manoeuvring can introduce additional challenges, while they
are implementing under state and parameter estimation
algorithms. This is due to the respective nonlinear
hydrodynamic forces and moments that can be a part of such
dynamic models and that those forces and moments cannot be
measured by on-board sensors. Various unobservable states and
parameters can be a part of such dynamic models in ship
manoeuvring. Therefore, the estimation algorithms may not able
to capture the respective system model due a large number of
unobservable states and parameters. On the other hand, the
accuracy and the system-model uncertainties of such models are
yet to be discovered. The eventual outcome is the divergence in
the parameter estimation process and the errors in the estimated
vessel states and parameters. Therefore, the prediction accuracy
of ship manoeuvres can be degraded.
A kinematic model-based ship manoeuvring prediction
algorithm is considered in this section. The modelling
difficulties in ship manoeuvring, i.e. complex vessel dynamic
conditions, can be avoided by this method. Similarly,
unobservable vessel states and parameters, i.e. forces and
moments, can also be avoided by this method and that can be
the main advantage in this approach. The hydrodynamic forces
and moments can be observed as vessel accelerations as an
important part of a kinematic model in ship manoeuvring. One
should also note that the vessel accelerations can be measured
accurately and that can improve the state and parameter
estimation process.
The ship on-board sensor data with a kinematic vessel
manoeuvring model and pivot point motion information has
been used to estimate the future vessel position and orientation
(i.e. heading) within a shorter time period. This approach has
been investigated in [10], where its capabilities are presented in
a simulated environment. This consists of a simplified
mathematical framework that is presented in Figure 4. The
current vessel position is presented by Pg (xg(t), yg(t)) with the
course speed vector of Vg(t) and the heading angle of ψg(t). In
addition, the future vessel position is presented by Pf (xf(t), yf(t))
with the course speed vector of Vf(t) and the heading angle of
ψf(t). One should note that the pivot points of the current and
future vessel pivot positions are denoted by Pgp and Pfg,
respectively. It has been shown that the current vessel position
and orientation (i.e. heading), measured by on-board sensor
measurements with an appropriate mathematical algorithm, can
be used to estimate the future vessel position and orientation
[10]. The algorithm is briefly discussed in this section and
consists of two sections.
The first part of this algorithm consists of estimating
the current vessel states and parameters by considering a
kinematic vessel manoeuvring model. This state and parameter
estimation process is supported by an extended Kalman filter
7 Copyright © 2019 by ASME
Fig. 4. Ship predictor in local scale.
(EKF) with the sensor measurements of vessel position,
heading, yaw rate and acceleration values. The second part of
this algorithm consists of estimating the future vessel states and
parameters by considering the current vessel states and
parameters. This state and parameter estimation process is
supported by a navigation vector dot and cross product
approach that consists of the pivot point information. Therefore,
the outcome of this algorithm is the future vessel position and
orientation (i.e. heading) within a short time interval. It is
expected that the respective weather and environmental
conditions can be observed as a part of ship motions under the
on-board sensors, i.e. wind and wave conditions.
While vessels are making circular type manoeuvres,
that can be associated with constant state and parameter values,
i.e. Rate of Turn (ROT), etc. as presented in Figure 4.
However, the prefect circle type manoeuvres cannot be possible
in some situations, due to external environmental conditions
that can influence on vessel states and parameters. Therefore,
the circular type ship manoeuvres can be resulted in parabolic
shapes due to external forces and moments. One should note
such external forces and moments are resulted due to weather
and environmental conditions and that can be slow varying
processes. Hence, the constant ROT values can be changed due
to external forces and moments in ship manoeuvring. Such
variations should be captured by the estimation algorithms to
predict vessel behaviour. This is where the EKF algorithm with
the proposed kinematic mathematical model for vessel
manoeuvres can play an important role by capturing such slight
variations in vessel states and parameters. That can improve the
predictability of ship manoeuvring, i.e. future vessel positions
and orientations. Since weather and environmental conditions
can be captured by the kinematic ship manoeuvring model and
estimation algorithm, this can be considered as another
advantage in this method. The respective computational results
for ship predictor in local scale is presented in Figure 5. That
consists of the current and future vessel positions and
orientations, predicted vessel pivot point trajectory and
predicted vessel trajectory. The respective data set is simulated
to verify the capabilities of the proposed approach as a part of
the advanced ship predictor in a local scale.
Global Scale
Ship behaviour can also be predicted on a global scale. This
entails predicting the future vessel position and orientation (i.e.
heading) within a longer time period. Based on the predicted
vessel behaviour, the ships in the respective encounter situation
can make necessary collision avoidance actions far in advance.
Simple speed or course alterations will likely be sufficient at
this point to avoid a close-encounter situation from occurring at
all, significantly reducing the risk associated with ship
operations. That can be the main advantage of having a ship
predictor in a global scale.
In general, such global behaviour predictions are
however difficult to conduct, as the future intentions of the
vessels are unknown in a majority of the encounter situations.
Historical AIS data however provide insight into ship behaviour
for specific regions. By exploiting these data sets in an
intelligent manner, one can estimate the future behaviour of a
selected vessel based on its own trajectory and past trajectories
of other vessels in the same region on a global scale. [11]
presents a thorough survey of methods to exploit AIS data for
ship navigation under such situations. The majority of the work
utilizing AIS data has focused on general traffic trends, anomaly
detection and long-term predictions [12-14]. Such information
can be very useful for general situation awareness, but limited
with respect to collision avoidance [15, 16]. Hence,
8 Copyright © 2019 by ASME
short-term vessel trajectory prediction techniques than can be
useful in terms of collision avoidance in such situations.
[17] builds upon the single point neighbour search
method (SPNSM) in [15] and introduces an improved multiple
trajectory extraction method (MTEM). In this method,
respective trajectories are extracted from a circular initial
cluster centered around the position of a selected vessel. The
trajectories with speed and heading values outside of a selected
threshold in the initial trajectory cluster are discarded in this
approach. This results in the extraction of trajectories with a
high degree of similarity to that of the selected vessel. The
clustering based iterative prediction technique [15] is then run
on the extracted trajectories for a given prediction horizon as
the next step. The resultant predicted vessel trajectory can
provide a higher degree of situation awareness of the ship
encounter situation in a global scale, as it can now have an
indication of the future intentions of the respective vessels. The
multiple extracted trajectories can also give an indication of the
spread of data for a given prediction horizon, where the
respective positions after a period of time corresponding the
desired prediction horizon can be described by a probability
density distribution. That represents the probability that the
selected vessel can be located after the selected period of time.
An example of predicting vessel positions using the MTEM is
visualized in Figure 6. A 30 minute trajectory prediction using
the SPNSM is compared with the MTEM indicating improved
performance for the MTEM in this figure. The orange contours
represent a kernel density estimate of the extracted trajectory
data after 30 minutes using Gaussian kernels.
Implementation Steps
As described previously, the advanced ship predictor can be
used to estimate the future vessel position with a short (i.e. in a
local scale) and long (i.e. in a global scale) time horizons.
Therefore, this can be a good supporting tool to estimate the
respective collision risk under various ship encounter situations.
This includes remote-controlled, autonomous, and manned ship
encounters, where human and system decision-making
situations can be encountered. The COLREGs and local
navigation rules and regulatory failures due to ship under-
actuation and human-system decision-making are expected as
the outcome of such situations. This should further be
investigated under both bridge simulator and realistic ocean-
going conditions of vessel encounters to support situation
awareness. The required modifications into the COLREGs and
local navigation rules and regulations [18] due to the respective
regulatory failures should also be investigated and that can
further enhance the required safety levels of future vessels.
The proposed advanced ship predictor, i.e. the ASPU,
is a part of the solution framework that has proposed to
overcome the respective challenges in situation awareness of
ship encounter situations of remote-controlled, autonomous, and
manned vessels, i.e. by introducing an appropriate collision risk
assessment procedure based on the advanced ship predictor.
Therefore, that should also be a part of the collision risk
assessment procedure. The ASPU can be evaluated under
bridge simulator conditions with various ship encounters [19],
initially and then onboard vessels under realistic ocean-going
conditions. Development and demonstration of the advanced
ship predictor to support situation awareness in ship navigation
and overcome the respective collision avoidance challenges in a
mixed environment can be the main contribution of this study.
That also introducs an appropriate collision risk assessment
methodology, while respecting the COLREGs and local
navigation rules and regulations and investigating additionally
required rule and regulation modifications.
9 Copyright © 2019 by ASME
System Perspective
The collision risk assessment methodology can also play an
important role in this proposed solution framework. The
collision risk assessment methods proposed by the research
community [20-24] should be further investigated and
appropriate techniques should be extracted towards the
proposed CRAU (see Figure 1 and 2). However, that should
accommodate both time and distance-based risk assessment
techniques to further improve the existing procedures, as
mentioned before. The CRAU with such methods can further
enhance the ASPU, i.e. the advanced ship predictor. It is
expected that vessel behaviour not only due to ship under-
actuation but also to navigator’s actions can also be reflected on
the ASPU, i.e. supported by both on-board sensor and AIS data.
The CRAU can initially be developed for a two-vessel
encounter situation and then that can be extrapolated for a
multi-vessel encounter situation. Furthermore, that should
initially be evaluated under computer simulations, then that can
be deployed towards bridge simulators in a later stage and
finally in realistic ocean-going conditions [25].
The SAM should be developed as a software module.
The facilities of the SAM can be further enhanced by
introducing adequate decision supporting features. The SAM
can be used as a decision supporting tool for remote-controlled
and manned vessels. It is expected that an advanced version of
the SAM should be developed with the PDMM and SAFM and
implemented as a system decisionaction execution model for
autonomous vessels. Furthermore, the COLREGs and other
local navigation rules and regulations can also be incorporated
into these modules to enhance its facilities towards decision-
making in autonomous ship navigation, as mentioned before.
Therefore, two versions of the SAM should be considered. The
first version should support manned and remote-controlled
vessels as a decision support unit and the second version
(advanced version) should support autonomous vessels as a
decision-making unit.
The SAM can be developed in bridge simulators to
evaluate its performance under remote-controlled, autonomous,
and manned ship encounter situations, as described previously
(see Figure 2). One version of the SAM is implemented as a
decision support system for remote-controlled and manned
vessels and another version (advanced version) is implemented
as a decision-making system for an autonomous vessel. A
considerable amount of complex ship encounters can be created
in this environment to evaluate the performance of the SAM. It
is also expected that the outcome of bridge simulator
experiments can consist of various collision and near-miss
situations, where possible regulatory failures can occur. Hence,
the respective regulatory failure situations and lessons learned
with possible reasons should be documented during these
experiments to further enhance the SAM.
The SAM can also be developed in integrated bridge
systems in ocean going vessels to evaluate its performance in a
mixed environment as described previously. Similarly, both
versions of the SAM can be implemented as a decision support
system for remote-controlled and manned vessels and as a
decision-making system for autonomous vessels, respectively. A
considerable amount of complex ship encounters can occur in
this environment, while evaluating the performance of the
SAM. It is also expected that the outcome of realistic ocean-
going vessel experiments can consist of various near-miss
situations (i.e. collision should be avoided at any cost), where
possible regulatory failures can occur. Hence, the respective
regulatory failure situations and lessons learned with possible
reasons should be documented during these experiments to
further enhance the SAM.
The respective regulatory failures observed in these
experimental conditions should further be studied. The outcome
can be used to improve the SAM and the improved SAM can be
used in experimental environments with additional experiments
to re-evaluate the outcome. Especially, the decisionaction
execution model (i.e. PDMM and SAFM) that has been adopted
for autonomous ship navigation can also be re-evaluated, and
appropriate modifications should be introduced during these
experiments. On the other hand, the same outcomes, i.e.
regulatory failures, can be considered to re-evaluate the
COLREGs and other local navigation rules and regulations in
both human and system perspectives. The required regulatory
modifications can also be investigated under the guidance of the
respective authorities.
CONCLUSIONS
This study considers a solution framework to support situation
awareness in a mixed environment, where remote-controlled,
autonomous and manned vessels are interacting. This consists
of an advanced ship predictor as a part of the collision risk
assessment methodology to identify unexpected vessel behavior
due to ship under-actuation, i.e. influenced by navigators’
actions and environmental conditions. The advanced ship
predictor consists of both local and global capabilities to
estimate vessel positions and orientations. Hence, situation
awareness in ship navigation with decision support (i.e. in
manned vessels) and decision making (i.e. in autonomous
vessels) facilities [26, 27] can be improved. However, this
should be evaluated under both bridge simulator and realistic
ocean-going conditions, where its capabilities in avoiding
collision and near-miss situations in remote-controlled,
autonomous and manned ship encounters can be experimented
with. The knowledge that can be created under such
experiments can enhance the situation awareness in a mixed
environment. The identification of the regulatory failures in
such ship encounter situations is also an important part of the
same knowledge, where possible modifications on the
COLREGs and local navigation rule and regulations can be
proposed. Finally, the documentation of the knowledge that can
be used to improve situation awareness with decision support
and decision making features, i.e. including the modified
10 Copyright © 2019 by ASME
COLREGs and local navigation rule and regulations, in future
ship encounters in a mixed environment is the ultimate
objective of this study.
ACKNOWLEDGMENTS
This work has been conducted under an internal funded project
of UiT The Arctic University of Norway- that supports towards
autonomous ship operations in the arctic region. This work is
also supported by the MARKOM2020 project, a development
project for maritime competence established by the Norwegian
Ministry of Education and Research in cooperation with the
Norwegian Ministry of Trade and Industry. The authors would
also like to express gratitude to the Norwegian Coastal
Administration for providing access to their AIS database.
REFERENCES
[1] L.P. Perera, "Autonomous Ship Navigation under Deep Learning
and the challenges in COLREGs," In Proceedings of the 37th
International Conference on Ocean, Offshore and Arctic
Engineering (OMAE 2018), Madrid, Spain, June, 2018
(OMAE2018-77672).
[2] IMO, “Convention on the International Regulations for
Preventing Collisions at Sea (COLREGs),” 1972.
[3] L.P. Perera, P. Oliveira and C. Guedes Soares, "System
Identification of Vessel Steering with Unstructured Uncertainties
by Persistent Excitation Maneuvers," IEEE Journal of Oceanic
Engineering, vol. 41, no. 3, 2016, pp. 515-528.
[4] Kalman, R. E. (1960). "A New Approach to Linear Filtering and
Prediction Problems". Journal of Basic Engineering. 82: 35.
[5] W. Xiaoyang, W. Yang, and Y. Xinping. Processes in Research
Methods for Modeling Accidents, China Safety Science Journal,
Vol. 26 No. 3, pp. 46-52.
[6] L.P. Perera, J.P. Carvalho and C. Guedes Soares, "Intelligent
ocean navigation & Fuzzy-Bayesian decision-action formulation,"
IEEE Journal of Oceanic Engineering, vol 37, no 2, 2012, pp
204-219.
[7] L.P. Perera, V. Ferrari, F.P. Santos, M.A. Hinostroza, and C.
Guedes Soares, "Experimental Evaluations on Ship Autonomous
Navigation & Collision Avoidance by Intelligent Guidance,"
IEEE Journal of Oceanic Engineering, vol. 40, no. 2, 2015, pp
374-387.
[8] T. A. Johansen, T. Perez and A. Cristofaro, "Ship Collision
Avoidance and COLREGS Compliance Using Simulation-Based
Control Behavior Selection with Predictive Hazard Assessment,"
in IEEE Transactions on Intelligent Transportation Systems, vol.
17, no. 12, pp. 3407-3422, Dec. 2016.
[9] L.P. Perera, J.P. Carvalho and C. Guedes Soares, "Solutions to the
Failures and Limitations of Mamdani Fuzzy Inference in Ship
Navigation," IEEE Transactions on Vehicular Technology, vol.
63, no. 4, 2014, pp 1539-1554.
[10] L.P. Perera, "Navigation Vector based Ship Manoeuvring
Prediction," Journal of Ocean Engineering, vol. 138, 2017, pp.
151-160.
[11] E. Tu, G. Zhang, L. Rachmawati, E. Rajabally, and G.-B. Huang,
“Exploiting AIS Data for Intelligent Maritime Navigation: A
Comprehensive Survey From Data to Methodology,” IEEE Trans.
Intell. Transp. Syst., pp. 124, 2017.
[12] K. Gunnar Aarsæther and T. Moan, “Estimating Navigation
Patterns from AIS,” J. Navig., vol. 62, no. 4, p. 587, Oct. 2009.
[13] G. Pallotta, M. Vespe, and K. Bryan, “Vessel Pattern Knowledge
Discovery from AIS Data: A Framework for Anomaly Detection
and Route Prediction,” Entropy, vol. 15, no. 12, pp. 2218 2245,
Jun. 2013.
[14] F. Mazzarella, V. F. Arguedas, and M. Vespe, “Knowledge-based
vessel position prediction using historical AIS data,” in 2015
Sensor Data Fusion: Trends, Solutions, Applications (SDF),
2015, pp. 16.
[15] S. Hexeberg, A. L. Flaten, B.-O. H. Eriksen, and E. F. Brekke,
“AIS-based vessel trajectory prediction,” in 2017 20th
International Conference on Information Fusion (Fusion), 2017,
pp. 18.
[16] B. R. Dalsnes, S. Hexeberg, A. L. Flaten, B.-O. H. Eriksen, and
E. F. Brekke, “The Neighbor Course Distribution Method with
Gaussian Mixture Models for AIS-Based Vessel Trajectory
Prediction,” in 2018 21st International Conference on
Information Fusion (FUSION), 2018, pp. 580587.
[17] B. Murray and L.P. Perera, "A Data-Driven Approach to Vessel
Trajectory Prediction for Safe Autonomous Ship Operations," In
Proceedings of the 13th International Conference on Digital
Information Management (ICDIM 2018), Berlin, Germany,
September, 2018.
[18] L.P. Perera, and C. Guedes Soares, "Collision Risk Detection and
Quantification in Ship Navigation with Integrated Bridge
Systems," Journal of Ocean Engineering, vol. 109, 2015, pp. 344-
354.
[19] P. Silveira, A. Teixeira, & C. Soares, (2013). Use of AIS Data to
Characterise Marine Traffic Patterns and Ship Collision Risk off
the Coast of Portugal. Journal of Navigation, 66(6), 879-898.
[20] Y.J. Xie, Q. Meng, Analysis of AIS-based Ship Traffic in the
Singapore Strait, Proceedings of the Inaugural World Transport
Convention, Beijing, China, June 4-6, 2017.
[21] T. A. Johansen, T. Perez and A. Cristofaro, "Ship Collision
Avoidance and COLREGS Compliance Using Simulation-Based
Control Behavior Selection With Predictive Hazard Assessment,"
in IEEE Transactions on Intelligent Transportation Systems, vol.
17, no. 12, pp. 3407-3422, Dec. 2016.
[22] W. Xiaoyang, W. Yang, E. Zio, Y. Xinping. What Kind of
Resilience Is Needed for Water Transport-A Lesson Learned from
Eastern Star. Safety Science. (Revised)
[23] R. Szlapczynski and J. Szlapczynska, "Review of ship safety
domains: Models and applications, "” Ocean Engineering, vol.
145, 2017, pp. 277-289.
[24] R. Szlapczynski and P. Krata, "Determining and visualizing safe
motion parameters of a ship navigating in severe weather
conditions, " Ocean Engineering, vol. 158, 2018, pp. 263-274.
[25] L.P. Perera, and C. Guedes Soares, "Collision Risk Detection and
Quantification in Ship Navigation with Integrated Bridge
Systems," Journal of Ocean Engineering, vol. 109, 2015, pp. 344-
354.
[26] L.P. Perera and B. Mo "Ship Performance and Navigation Data
Compression and Communication under Autoencoder System
Architecture," Journal of Ocean Engineering and Science, 2018
(DIO: 10.1016/j.joes.2018.04.002).
[27] L.P. Perera and B. Mo, "Machine Intelligence based Data
Handling Framework for Ship Energy Efficiency," IEEE
Transactions on Vehicular Technology, vol. 66, no. 10, 2017, pp.
8659-8666.
... Though dynamic motion models are preferable in many maritime applications, they can have some shortcomings. One of the shortcomings is the difficulty of applying the dynamic motion models to vessel state estimation [9]. This is because the nonlinear hydrodynamic forces and moments required by the dynamic motion models cannot be measured by on-board sensors. ...
... Another shortcoming is that the existing dynamic motion models are still insufficient when dealing with vessels' nonlinear characteristics [9]. Underactuation is a typical nonlinear property of ship maneuvering when vessels are equipped with a single rudder-propeller-thrust system as the power input unit. ...
Conference Paper
Full-text available
Autonomous shipping with adequate decision support systems is widely considered as a high-potential development direction in the maritime industry in the upcoming years. Prediction technologies are one of the key components in these decision support systems and they usually require a large number of data sets to estimate vessel states. Certain vessel motion models are generally implemented with the above-mentioned prediction technologies to improve the accuracy and robustness of the estimated states. In contrast to wider research studies of different motion models for the applications of ground vehicles, the studies of appropriate motion models for maritime transport applications are still insufficient. Therefore, it is necessary to develop reliable motion models for vessels, and that can improve the decision supporting capabilities in future vessels, especially in autonomous shipping. In this paper, two kinematic motion models which can be used to estimate various vessel maneuvering states are examined and compared. In the current stage, the proposed models are used to investigate ship maneuvers produced by a ship bridge simulator. Two nonlinear filter algorithms combined with Monte Carlo-based simulation tests are applied to estimate the respective vessel states. In the conclusion, a comprehensive comparison of the estimation algorithms is presented with the estimated vessel states. Hence, this study provides robust and convenient estimation algorithms that can support autonomous shipping navigation in the future.
... In existing studies, ship behavior generally refers to behavior such as path, speed, and course, etc. A solution framework to support situation awareness in a mixed environment is considered by Perera et al. [12]. The review on ship behavior established a generic behavior identification model, but which did not consider the behavioral differences of different ship types under specific external environmental constraints [13]. ...
Article
Full-text available
Through the continuous development of intellectualization, considering the lifecycle of ships, the future of a waterborne traffic system is bound to be a mixed scenario where intelligent ships of different autonomy levels co-exist, i.e., mixed waterborne traffic. According to the three modules of ships’ perception, decision-making, and execution, the roles of humans and machines under different autonomy levels are analyzed. This paper analyzes and summarizes the intelligent algorithms related to the three modules proposed in the last five years. Starting from the characteristics of the algorithms, the behavior characteristics of ships with different autonomous levels are analyzed. The results show that in terms of information perception, relying on the information perception techniques and risk analysis methods, the ship situation can be judged, and the collision risk is evaluated. The risk can be expressed in two forms, being graphical and numerical. The graphical images intuitively present the risk level, while the numerical results are easier to apply into the control link of ships. In the future, it could be considered to establish a risk perception system with digital and visual integration, which will be more efficient and accurate in risk identification. With respect to intelligent decision-making, currently, unmanned ships mostly use intelligent algorithms to make decisions and tend to achieve both safe and efficient collision avoidance goals in a high-complexity manner. Finally, regarding execution, the advanced power control devices could improve the ship’s maneuverability, and the motion control algorithms help to achieve the real-time control of the ship’s motion state, so as to further improve the speed and accuracy of ship motion control. With the upgrading of the autonomy level, the ship’s behavior develops in a safer, more efficient, and more environment-friendly manner.
... However, such techniques will not be useful for prediction horizons greater than a few minutes. Perera and Murray [11] suggested to introduce an advanced ship predictor to aid maritime situation awareness. The predictor is comprised of a local and global predictor to overcome such issues. ...
Article
Full-text available
This study presents a deep learning framework to support regional ship behavior prediction using historical AIS data. The framework is meant to aid in proactive collision avoidance, in order to enhance the safety of maritime transportation systems. In this study, it is suggested to decompose the historical ship behavior in a given geographical region into clusters. Each cluster will contain trajectories with similar behavior characteristics. For each unique cluster, the method generates a local model to describe the local behavior in the cluster. In this manner, higher fidelity predictions can be facilitated compared to training a model on all available historical behavior. The study suggests to cluster historical trajectories using a variational recurrent autoencoder and the Hierarchical Density-Based Spatial Clustering of Applications with Noise algorithm. The past behavior of a selected vessel is then classified to the most likely clusters of behavior based on the softmax distribution. Each local model consists of a sequence-to-sequence model with attention. When utilizing the deep learning framework, a user inputs the past trajectory of a selected vessel, and the framework outputs the most likely future trajectories. The model was evaluated using a geographical region as a test case, with successful results.
... Vessel dynamic models usually contain nonlinear factors, such as hydrodynamic forces and moments. These nonlinear forces and moments are difficult to measure or observe by onboard sensors (Perera and Murray, 2019). Without enough measurements, estimation algorithms may not guarantee converged results. ...
Conference Paper
Full-text available
The vessel state and parameter estimation is essential to ship maneuvering and collision avoidance. This study presents an application of a particle filter algorithm to estimate vessel states and parameters. Particularly, to capture the impact of the vessel underactuated property and complex environmental-induced disturbances, the estimation process contains a kinematic curvilinear motion model that describes vessel motions. The estimated result can help ship navigators or onboard computers to well comprehend present vessel maneuvering conditions. Besides, it can also serve as a necessary data source for future trajectory predictions for ocean-going vessels. Therefore, it can be integrated into situation awareness type applications in vessels that can improve the navigation safety for both manned and autonomous vessels.
... Sensor function is critical for situation awareness in a SAR operation. It is enabled by the collection and integration of information from on-board sensors (e.g., heat, sonar and sound detection sensors), cameras and the automatic identification system (AIS) through satellites (Perera and Murray 2019). Even the lookout requirement under the International Regulations for Preventing Collisions at Sea (COLREG) (COLREG 1972, rule 5) is expected to be fulfilled by sensor technology (Lloyd's Register 2017 Code, chap 4, s 4.1.5; ...
Article
Maritime Autonomous Surface Ships (MASS) have obtained much attention in recent years. Navigation with human–machine cooperation is the core for L1–L3 MASS. In this article, the development and trend of MASS are collected, and main research institutions are analysed based on MASS experimental projects and publications. The navigation modes of MASS are given, and main topics involved in human–machine cooperative navigation are presented. We focus on the research progress in the four aspects, i.e., modelling of ship navigating behaviour, enhanced navigation situation awareness, human–machine cooperation decision-making and human–machine cooperation navigation control, and then we compare the different methods and technologies for different issues of human–machine cooperative navigation. On the basis of current development, the deficiencies and unsolved problems of MASS human–machine cooperative navigation are analysed. The key theories and technologies of identification and risk assessment of navigating behaviour, situation awareness enhancement, human–machine cooperative decision-making, human–machine cooperative navigation, experimental platform and a case study of human–machine cooperative navigation are investigated, and relevant feasible approaches are provided. We believe the human–machine cooperative navigation will be an important breakthrough of MASS development, and inland and offshore waters will be the suitable navigation area for testing the performance of ship human–machine cooperative navigation.
Article
Full-text available
Autonomous ship technology is developing at a rapid pace, with the aim of facilitating safe ship operations. Collision avoidance is one of the most critical tasks that autonomous ships must handle. To support the level of safety associated with collision avoidance, this study suggests to provide autonomous ships with the ability to conduct proactive collision avoidance maneuvers. Proactive collision avoidance entails predicting future encounter situations, such that they can preemptively be avoided. However, any such actions must adhere to relevant navigation rules and regulations. As such, it is suggested to predict encounter situations far in advance, i.e. prior to risk of collision existing. Any actions can, therefore, be conducted prior to the applicability of the COLREGs. As such, simple corrective measures, e.g. minor speed and/or heading alterations, can prevent close encounter situations from arising, reducing the overall risk associated with autonomous ship operations, as well as improving traffic flow. This study suggests to facilitate this ability by emulating the development of situation awareness in ship navigators through machine learning. By leveraging historical AIS data to serve as artificial navigational experience, long-range trajectory predictions can be facilitated in a similar manner those conducted by human navigators, where such predictions provide the basis for proactive collision avoidance actions. The development of human situation awareness is, therefore, presented, and relevant machine learning techniques are discussed to emulate the same mechanisms.
Article
The multi-ship encounter situation at sea is characterized by high complexity and uncertainty, which is a big challenge for both traditional ships and the new autonomous ships. In order to make reasonable navigation decisions and perform well under multi-ship encounter situation, it is necessary to grasp the current scenario correctly and intelligently. Therefore, in this paper, an adaptive understanding model for multi-ship encounter situation is proposed. The core function of this model is to infer the navigation intention of other target ships under the same situation. This model is mainly composed of two sub-models. One is the ship encounter situation analysis model, which realizes the cognition of the whole encounter scenario from the global perspective by maintaining the “double matrix”. The second is the ship navigation intention inference model, the key part of the model is a set of well-designed fuzzy inference system. The output of the encounter situation analysis model is the input of the intention inference model, and these two models are closely linked to form a unified whole. This model is verified by both simulation-based and real scenario-based experiments, the results show that this model can perform well under the complex multi-ship encounter situation. Moreover, some necessary discussion and analysis for this inference model are also stated at the end of this paper, in the future, we expect that this model can be applied in real situations.
Conference Paper
Full-text available
Autonomous ship technology is developing at a rapid pace, with the aim of facilitating safe ship operations. Collision avoidance is one of the most critical tasks that autonomous ships must handle. To support the level of safety associated with collision avoidance, this study suggests to provide autonomous ships with the ability to conduct proactive collision avoidance maneuvers. Proactive collision avoidance entails predicting future encounter situations, such that they can preemptively be avoided. However, any such actions must adhere to relevant navigation rules and regulations. As such, it is suggested to predict encounter situations far in advance, i.e. prior to risk of collision existing. Any actions can, therefore, be conducted prior to the applicability of the COLREGs. As such, simple corrective measures, e.g. minor speed and/or heading alterations, can prevent close encounter situations from arising, reducing the overall risk associated with autonomous ship operations, as well as improving traffic flow. This study suggests to facilitate this ability by emulating the development of situation awareness in ship navigators through machine learning. By leveraging historical AIS data to serve as artificial navigational experience, long-range trajectory predictions can be facilitated in a similar manner those conducted by human navigators, where such predictions provide the basis for proactive collision avoidance actions. The development of human situation awareness is, therefore, presented, and relevant machine learning techniques are discussed to emulate the same mechanisms.
Article
To improve the safety and efficiency when ships pass through a lock, a Cooperative Ship Formation System (CSFS) in the lock waterway is proposed. An improved ship formation structure with the combination of leader–follower and behaviour-based structure is designed, and the relevant control modes are described according to the process of ship formation passing through the lock. MilliMetre-Wave (MMW) radars are introduced for the perception of ship states. Moreover, speed control, stop control and distance keeping control methods are proposed, which are related to the distance keeping control mode, speed control mode, stop control mode, stop-finished mode and emergency stop mode. Finally, a real CSFS platform with three different engineering ships was implemented in the Gezhouba Lock waterway to verify the effectiveness of CSFS and proposed control methods. Several experimental tests were carried out with the CSFS platform. The experiment results show that: 1) the average tracking error of the distance keeping control is 2.43 m; 2) the average speed control error of the leading ship is 0.08 m/s; 3) the leading ship can stop at around 2.0 m to the stop line, and both the leading ship and the following ship can switch smoothly between the different control modes.
Conference Paper
Full-text available
A general framework to support the navigation side of autonomous ships is discussed in this study. That consists of various maritime technologies to achieve the required level of ocean autonomy. Decision-making processes in autonomous vessels will play an important role under such ocean autonomy, therefore the same technologies should consist of adequate system intelligence. Each onboard application in autonomous vessels may require localized decision-making modules, therefore that will introduce a distributed intelligence type strategy. Hence, future ships will be agent-based systems with distributed intelligence throughout vessels. The main core of this agent should consist of deep learning type technology that has presented promising results in other transportation systems, i.e. self-driving cars. Deep learning can capture helmsman behavior, therefore that type system intelligence can be used to navigate autonomous vessels. Furthermore, additional decision supporting layers should also be developed to facilitate deep learning type technology including situation awareness and collision avoidance. Ship collision avoidance is regulated by the Convention on the International Regulations for Preventing Collisions at Sea, 1972 (COLREGs) under open sea areas. Hence, a general overview of the COLREGs and its implementation challenges, i.e. regulatory failures and violations, under autonomous ships are also discussed with the possible solutions as the main contribution of this study. Furthermore, additional considerations, i.e. performance standards with the applicable limits of liability, terms, expectations and conditions, towards evaluating ship behavior as an agent-based system on collision avoidance situations are also illustrated in this study.
Article
Full-text available
The Automatic Identification System (AIS) tracks vessel movement by means of electronic exchange of navigation data between vessels, with onboard transceiver, terrestrial and/or satellite base stations. The gathered data contains a wealth of information useful for maritime safety, security and efficiency. This paper surveys AIS data sources and relevant aspects of navigation in which such data is or could be exploited for safety of seafaring, namely traffic anomaly detection, route estimation, collision prediction and path planning.
Conference Paper
Autonomous vehicles will be an integral part of future transportation systems, and the maritime industry is working towards developing methods to ensure safe autonomous ship operations. One of the major challenges in realizing autonomous ships is ensuring effective collision avoidance technologies. Autonomous vessels must have a higher degree of situation awareness to detect other vessels, predict their future intentions, and evaluate the respective collision risk. One step in achieving this goal is to predict other vessel trajectories accurately. In this paper, a data-driven approach to vessel trajectory prediction for time horizons of 5-30 minutes utilizing historical AIS data is evaluated. A clustering based Single Point Neighbor Search Method is investigated along with a novel Multiple Trajectory Extraction Method. Predictions have been conducted using these methods and compared with the Constant Velocity Method. Additionally, the Multiple Trajectory Extraction Method is utilized to evaluate estimated ship routes.
Article
https://authors.elsevier.com/a/1Wum86nh6ooro Free access till June 05, 2018 The paper presents a method of determining, organizing and displaying ship collision avoidance information, which is based on the Collision Threat Parameters Area (CTPA) technique. The method makes it possible to visualize navigational threats as well as possible collision avoidance manoeuvres. The solution is focused on supporting navigation in severe weather conditions. Normally collision avoidance decisions are made taking into account targets' motion parameters, International Regulations for Preventing Collisions at Sea (COLREGS) and navigational obstacles. However, in hard weather conditions each manoeuvre has to be additionally checked to assess whether it is safe in terms of ship's stability. Therefore the proposed method provides four types of information: motion parameters of targets within a given range, combinations of own course and speed which collide with those targets, combinations of own course and speed which would lead to grounding within a specified time and combinations of own course and speed, which could result in stability-related dynamical threats. Optionally it is also possible to display only manoeuvres compliant with COLREGS. A superposition of these types of data enables a navigator to choose an efficient manoeuvre in a situation when possibilities are limited by weather conditions and actual characteristics of the ship stability. Free access till June 05, 2018 https://authors.elsevier.com/a/1Wum86nh6ooro
Article
Free access to the article valid for 50 days, until November 11, 2017: https://authors.elsevier.com/a/1Vm796nh6k-4p Ship safety domain is a term which is widely used in research on collision avoidance and traffic engineering among others. Classic ship domains have been compared in multiple reports. However, up till now there has been no work summing up contemporary research in this field. The paper offers a systematic and critical review of the newer ship domain models and related research. It discusses multiple differences in approach to ship domain concept: from definitions and safety criteria, through research methodologies and factors taken into account, to sometimes largely different results obtained by various authors. The paper also points out some interpretation ambiguities related to ship domain and sums up present trends of its development and applications. Free access to the article valid for 50 days, until November 11, 2017: https://authors.elsevier.com/a/1Vm796nh6k-4p
Article
A novel mathematical framework for predicting ship maneuvers within a short time interval is presented in this study. The first part of this study consists of estimating the required vessel states and parameters by considering a kinematic vessel maneuvering model. That is supported by an extended Kalman filter (EKF), where vessel position, heading, yaw rate and acceleration measurements are used. Then, the estimated vessel states and parameters are used to derive the respective navigation vectors that consist the pivot point information. The second part of this study consists of predicting the future vessel position and orientation (i.e. heading) within a short time interval by a vector product based algorithm, where the respective navigation vectors are used. The main advantage in this method is that the proposed framework can accommodate external environmental conditions in ship navigation and that feature improves the predictability of vessel maneuvers. Finally, the proposed mathematical framework is simulated and successful computational results in predicting ship maneuvers are presented in this study. Therefore, that can be implemented in modern integrated bridge systems to improve the navigation safety in maritime transportation.
Article
Appropriate navigation strategies should be developed to overcome the current shipping industrial challenges under emission control based energy efficiency measures. Effective navigation strategies should be based on accurate ship performance and navigation information, therefore various onboard data handling systems are installed on ships to collect large-scale data sets. Ship performance and navigation data that are collected to develop such navigation strategies can be an integrated part of the ship energy efficiency management plan (SEEMP). Hence, SEEMP with various navigation strategies can play an important part of e-navigation under modern integrated bridge systems. This study proposes a machine intelligence (MI) based data handling framework for ship performance and navigation data to improve the quality of the respective navigation strategies. The prosed framework is divided into two main sections of pre and post processing. The data pre-processing is an onboard application that consists of sensor faults detection, data classification and data compression steps. The data post processing is a shore-based application (i.e. in data centers) and that consists of data expansion, integrity verification and data regression steps. Finally, a ship performance and navigation data set of a selected vessel is analyzed through the proposed framework and successful results are presented in this study.
Article
System identification of vessel steering associated with unstructured uncertainties is considered in this paper. The initial model of vessel steering is derived by a modified second-order Nomoto model (i.e., nonlinear vessel steering with stochastic state-parameter conditions). However, that model introduces various difficulties in system identification, due to the presence of a large number of states and parameters and system nonlinearities. Therefore, partial feedback linearization is proposed to simplify the proposed model, where the system-model unstructured uncertainties can also be separated. Furthermore, partial feedback linearization reduces the number of states and parameters and the system nonlinearities, given the resulting reduced-order state model. Then, the system identification approach is carried out, for both models (i.e., full state model and reduced-order state model), resorting to an extended Kalman filter (EKF). As illustrated in the results, the reduced-order model was able to successfully identify the required states and parameters when compared to the full state model in vessel steering under persistent excitation maneuvers. Therefore, the proposed approach can be used in a wide range of system identification applications.