Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Ergonomics

ISSN: 0014-0139 (Print) 1366-5847 (Online) Journal homepage: www.tandfonline.com/journals/terg20

The human element in future Maritime Operations


– perceived impact of autonomous shipping

Steven C. Mallam, Salman Nazir & Amit Sharma

To cite this article: Steven C. Mallam, Salman Nazir & Amit Sharma (2020) The human element
in future Maritime Operations – perceived impact of autonomous shipping, Ergonomics, 63:3,
334-345, DOI: 10.1080/00140139.2019.1659995

To link to this article: https://doi.org/10.1080/00140139.2019.1659995

Published online: 06 Sep 2019.

Submit your article to this journal

Article views: 3094

View related articles

View Crossmark data

Citing articles: 38 View citing articles

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=terg20
ERGONOMICS
2020, VOL. 63, NO. 3, 334–345
https://doi.org/10.1080/00140139.2019.1659995

ARTICLE

The human element in future Maritime Operations – perceived impact of


autonomous shipping
Steven C. Mallam, Salman Nazir and Amit Sharma
Training and Assessment Research Group, Department of Maritime Operations, Faculty of Technology, Natural Sciences and Maritime
Sciences, University of South Eastern Norway, Borre, Norway

ABSTRACT ARTICLE HISTORY


The role of the human element within complex socio-technical systems is continually being Received 31 October 2018
transformed and redefined by technological advancement. Autonomous operations across vary- Accepted 15 August 2019
ing transport domains are in differing stages of realisation and practical implementation, and
KEYWORDS
specifically within maritime operations, is still in its infancy. This study explores the potential
Autonomous systems;
effects of autonomous technologies on future work organisation and roles of humans within digitisation; work
maritime operations. Ten Subject-Matter Experts working within industry and academia were organisation; trans-
interviewed to elicit their perspectives on the current state and future implications of autono- port; training
mous technologies. Four main themes emerged: (i) Trust, (ii) Awareness and Understanding, (iii)
Control, (iv) Training and Organisation of Work. A fuzzier fifth theme also appeared in the data
analysis: (v) Practical Implementation Considerations, which encompassed various sub-topics
related to real-world implementation of autonomous ships. The results provide a framework of
human element issues relevant for the organisation and implementation of autonomous mari-
time operations.

Practitioner summary: As autonomous shipping rapidly moves closer to real-world implementa-


tion, it is critical to develop an understanding of future roles of humans in autonomous maritime
operations. By eliciting expert knowledge from academics and practitioners, we establish a frame-
work of relevant issues facing humans in emerging autonomous systems and operations at sea.

Abbreviations: AVG: average; COLREGS: International Regulations for Preventing Collisions at


Sea; ECDIS: Electronic Chart Display and Information System; IMDG: International Maritime
Dangerous Goods Code; IMO: The International Maritime Organization; MASS: Maritime
Autonomous Surface Ships; MAX: maximum; MIN: minimum; PACT: Pilot Authorisation and
Control of Tasks; SD: standard deviation; SME: subject-matter experts; SOLAS: The International
Convention for the Safety of Life at Sea; STCW: The International Convention on Standards of
Training, Certification and Watchkeeping for Seafarers

1. Introduction coastal tourism operate in frequently harsh, isolated


and hazardous environments. As such, they have
Advancing technologies and digital transformation
developed as engineering and technology-rich
continue to alter how work tasks and systems are
domains, typically interested in applying new solutions
designed, organised and operated. As the sophistica-
that can enhance both safety and efficiency of opera-
tion and feasibility of implementing highly automated tions (Allen 2009; Stopford 2009). Over the past sev-
and autonomous technologies increases, the way in eral decades technological advancements in naval
which the human element interacts and contributes to engineering, computerisation and the codification of
achieve a system’s goals will continue to transform. international regulatory standards for ship design, con-
Thus, the role of the human, including their training, struction and operations have contributed towards
work tasks and competencies required for managing steadily increasing safety performance, reducing the
complex socio-technical systems are continuously likelihood of maritime accidents (Allianz (Allianz Global
redefined by technological advancements. Corporate & Specialty) 2012, 2018; Chauvin et al.
Maritime-related industries, such as shipping, off- 2013). The general trend of increased safety at sea
shore energy, commercial fishing, aquaculture and across the global fleet is occurring even whilst an

CONTACT Steven C. Mallam steven.mallam@usn.no Training and Assessment Research Group, Department of Maritime Operations, Faculty of
Technology, Natural Sciences and Maritime Sciences, University of South Eastern Norway, Borre, Norway
ß 2019 Informa UK Limited, trading as Taylor & Francis Group
ERGONOMICS 335

inverse relationship exists between ship sizes and 1.1. Purpose


crew numbers: larger and larger ships are being built,
The purpose of this paper is to explore the perceived
requiring relatively smaller numbers of onboard crew
impact and implications of autonomous technologies
to successfully operate them (Mallam 2016). This
on shipping operations and the human element in the
reduction of onboard crew is enabled by the increas-
maritime domain. This research is based on interview
ing reliability and efficacy of automated equipment
data from Subject-Matter Experts in order to elicit
and operations. Modern ships are able to receive and expert perspectives on how applications of autono-
transmit increasingly large amounts of data. While this mous systems can, and have already, impacted their
has enabled improved coordination and execution of work and system processes. Individuals from a broad
many maritime operations, it has also given rise to the spectrum of areas and competencies within the mari-
possibility for remote command and control of ships time domain were recruited. Two distinct cohorts
via shore-side control centres where land-based opera- emerged: academic researchers (e.g. human factors
tors monitor unmanned ships at sea (Rødseth and and ergonomics researchers working within academia)
Burmeister 2015). Furthermore, shipping companies and industry practitioners (e.g. former operators, engi-
and suppliers are now beginning to invest in the neers and computer scientists working within mari-
development of autonomously operated unmanned time-related and process industries) in order to better
ships. There are several commercial projects, such as understand emerging technologies’ perceived implica-
the Yara Birkeland, which aims to deliver an tions, benefits and challenges.
unmanned ship for short sea coastal commercial util-
isation by the end of the year 2020 (KM 2018).
Industry giants such as Wilhelmsen and Kongsberg
2. Background
have already announced the establishment of the 2.1. Defining automation and autonomy
world’s first autonomous shipping company aiming to
To illustrate the contemporary state of development in
address the complete value chain for autonomous
autonomous systems, it is useful to depict automation
shipping (Wilhelmsen 2018).
as a continuum and to clarify associated definitions. In
However, as new technologies and methods
its most rudimentary definition, ‘autonomy’ refers to
develop there is typically a delay between its initial
self-regulation or self-government. ‘Automation’ can be
creation, feasibility and adoption (Doraszelski 2004;
defined as the execution of a machine agent(s) (usually
Mallam, Lundh, and MacKinnon 2017). New technolo- computerized) of a function or task that was previously
gies can also have unintended and unforeseen conse- performed by a human (Parasuraman and Riley 1997).
quences that may even negatively affect a system, its Most of the agents in modern industrial applications
users and its processes. For example, numerous acci- fall somewhere between being completely manually
dents in safety-critical domains, including maritime operated (by humans) and completely autonomous (by
industries, have occurred not despite, but because of, technology). Therefore, in the context of work systems,
automation and technology (Lutzhoft and Dekker the level of ‘autonomy’ can be described by defining
2002; Nazir, Kluge, and Manca 2014). Humans are fun- the roles and relationships between humans and
damentally poor passive monitors of automated sys- machines in operations. There have been various mod-
tems (Parasuraman 1987), and issues such as loss of els proposed that describe the degrees of autonomy as
skills, out-of-loop syndrome, mental workload and a a function of who makes the decisions in complex
lack of trust in automated systems contribute to acci- socio-technical systems: human or machine (e.g.
dents (Bainbridge 1983; Endsley and Kiris 1995; Hoff Endsley and Kiris 1995; Parasuraman, Sheridan, and
and Bashir 2015; Kim, Nazir, and Øvergård 2016; Wickens 2000; Sheridan 1992). Figure 1 illustrates a
Strauch 2017). Thus, as emerging autonomous tech- five-level continuum of automation as defined by
nologies become increasingly viable, a better under- Endsley and Kiris (1995), ranging from an absence of
standing of how to facilitate these applications and automation (i.e. Level 0) through increasing levels until
the human element is not only new but also existing reaching full automation (i.e. Level 4), where a system
systems must be investigated. Merchant shipping is has complete autonomy to pursue its goals and exe-
conventionally classified as an industry slow to react cute desired actions.
to disruptive changes. It may, therefore, be beneficial Different transportation domains have adopted spe-
to take a proactive stance to facilitate smoother and cific frameworks for defining the levels of automation
safer transition to this new era of operations. in their operations. Although details, such as the
336 S. C. MALLAM ET AL.

Figure 1. Levels of automation in socio-technical systems (adapted from Endsley and Kiris 1995).

number of levels in a model or operational-specific Many operational functions relating to navigational


nomenclature may vary, similarities across differing tasks and equipment (e.g. bridge systems, such as
frameworks are apparent. For example, in the aviation autopilot, Electronic Chart Display and Information
sector, the Pilot Authorisation and Control of Tasks System [ECDIS], Dynamic Positioning, etc.) are now
(PACT) framework outlines the six different levels of digitised and highly automated, moving operations
automation that can exist between humans and the closer to automation levels 2, 3 and 4 of the Endsley
system (Bonner et al. 2000). These different levels and Kiris (1995) model.
demarcate control between humans and the system. Automation has allowed for the redistribution of
PACT defines these operational levels of an aircraft in operational duties between human and technical
one of the three modes for routine, as well as safety agents, resulting in ships operating with far less crew
critical events: (1) human command, (2) assisted and today than in the past (Grech, Horberry, and Koester
(3) automatic (Richards and Stedmon 2016). Within the 2008; Progoulaki and Roe 2011). Reducing the amount
automotive sector, the Society of Automotive of onboard crew is a relatively simple and straightfor-
Engineers has proposed six levels of driving automa- ward method for shipping companies to cut
tion for cars: (0) no automation, (1) driver assistance,
operational overhead costs (Allianz (Allianz Global
(2) partial automation, (3) conditional automation, (4)
Corporate & Specialty) 2017; Hummels 2007). For
high automation and (5) full automation (Society of
example, crew costs for bulk carriers over a five-year
Automotive Engineers 2018). It is expected that auto-
period averaged approximately 10% of total charter
mation levels 3 and 4 will be introduced by 2020, and
expenses, and in some instances up to 36% of total
by 2030 more than 50% of automobiles could have
costs (Rødseth and Burmeister 2012). Although
level 5 (full automation) features (Kyriakidis et al.
unmanned and autonomous ship operations will not
2019). The maritime domain has recently introduced
similar frameworks to cater for design, operation and completely reduce these operational expenses (e.g.
legislative issues arising with the recent developments some level of onshore control and/or monitoring by
and planned implementation of autonomous shipping humans will be necessary, as well as ship and equip-
(IMO 2018b). ment maintenance), it will reduce and reorganise the
human element within the system.
There are a myriad of potential societal benefits
2.2. Automation and autonomy in shipping with autonomous shipping, including: environmental
Though characterised as a generally conservative busi- advantages due to fuel savings, enhanced safety,
ness in comparison to other industries, merchant ship- increased cost efficiency and productivity, and opti-
ping has undergone vast transformations in order to mised logistical chains and infrastructure (DMA 2017).
remain competitive (Stopford 2009). The technologies Unmanned and autonomous shipping operations also
and operations of contemporary merchant ships stand contribute to the sustainability of maritime transport
in stark contrast to those of just a few decades earlier. as a whole, through increased economic, ecological
ERGONOMICS 337

Table 1. Degrees of autonomy of MASS (adapted from snowball sampling) were utilised to recruit participants
IMO 2018b). from the researcher’s network. Participants worked for
Degree of autonomy Description various academic institutions, government and private
Ship with automated Seafarers are on board to operate
processes and and control shipboard systems
sector companies in Northern Europe and the United
decision support and functions. Some operations States. All participants completed an informed consent
may be automated.
Remotely controlled ship with The ship is controlled and operated
form prior to participation in the interviews. This
seafarers on board from another location, but research project and its interview questions were reg-
seafarers are onboard.
Remotely controlled ship The ship is controlled and operated
istered and approved by the Norwegian Centre for
without seafarers on board from another location. There are Research Data (project no. 56374).
no seafarers onboard.
Fully autonomous ship The operating system of the ship is
able to make decisions and
determine actions by itself.
3.2. Procedure
and social sustainability (Rødseth and Burmeister The interviews followed a semi-structured question
2012). The International Maritime Organisation (IMO) script that focussed on three general areas: (1) the
has established a regulatory scope for what they call definition and concept of autonomy and autonomous
‘Maritime Autonomous Surface Ships’ (MASS), which systems, (2) perceived challenges, and (3) the role of
can operate independently of human interaction (IMO humans in future training and work paradigms (see
2018a). Similar to the Endsley and Kiris (1995) model Table 3). Six of the ten participant interviews were in
of generalised automation levels, the IMO has estab- English; four were in Norwegian (due to both inter-
lished degrees of autonomy for MASS operations viewer and interviewee being native speakers).
(see Table 1). These degrees are non-hierarchal and a Norwegian-spoken interviews were translated to
MASS can operate at one or more degrees of auton- English for data analysis. Participants were interviewed
omy throughout a single voyage (IMO 2018b). For individually and audio recorded.
example, a ship would use differing degrees of auton-
omy at different phases of a voyage (e.g. berthing vs.
port entry/departure vs. sailing in open waters)
(NFAS 2017). 3.3. Data analysis
All interviews were transcribed verbatim post-hoc from
3. Methods the audio recordings. The interview transcripts were
analysed using Thematic Analysis (Braun and Clarke
3.1. Participants 2006). The six-step process of Thematic Analysis was
Ten Subject-Matter Experts (SME) participants were used as a framework to code content and define com-
recruited for individual interview sessions. SMEs con- mon patterns within the data: (1) familiarisation with
sisted of (i) academic researchers (working experience the data; (2) generating initial codes; (3) searching for
in field: AVG ¼ 17.4 yrs; SD ¼ 5.0 yrs; MIN ¼ 10.0 yrs; themes among codes; (4) reviewing themes; (5) defin-
MAX ¼ 25.0 yrs), and (ii) industry practitioners (working ing and naming themes; (6) final reporting. Three
experience in field: AVG ¼ 25.0 yrs; SD ¼ 10.5 yrs; members of the research team first independently
MIN ¼ 5.0 yrs; MAX ¼ 35.0 yrs) working within various reviewed and analysed the interview transcripts. This
positions (e.g. researchers, training instructors, market- was followed by a workshop where the researchers
ing, managers, analysts) where automation and auton- presented their individual findings in order to validate
omy are influencing their areas of interest (see individual interpretations and analysis of the qualita-
Table 2). Non-probability sampling (convenience and tive data and establish final results.

Table 2. Participant backgrounds and expertise.


Discipline Current position/employer Domains of expertise
Academic researchers  3 Full Professors Safety Management, Design, Maritime Technology,
 1 Associate Professor Human Factors, Training
 1 Research Scientist
Industry practitioners  2 Training Centres Competence Development/Assurance, Maritime
 1 Equipment Manufacturer Transport, Simulation/Modelling, Human
 1 National Transport Agency Factors, Energy Sector
 1 Classification Society
338 S. C. MALLAM ET AL.

Table 3. Participant interview question transcript.


1 Introduction and definition questions:
1.1 What is your definition and concept of “autonomy”?
1.2 Could you please discuss an example(s) of advancing
automation and/or autonomy in your field?
2 Perceived challenges and changes questions:
2.1 What will be the biggest challenge faced in implementing
autonomous systems?
2.2 What will be the biggest challenge faced in operating
autonomous systems?
2.3 How will autonomy change every day processes/operations in
your field of interest?
2.4 How do you see the future impacts of autonomy in
your domain?
3 Role of human questions:
3.1 What is the new role of the human/operator in
autonomous systems?
3.2 How will increasing automation and autonomy affect the
training of new and existing workers?
3.3 What type of skills will be required for these emerging
technologies?
3.4 What solutions are needed in training?

Figure 2. Overview of themes that emerged from partici-


4. Results pant interviews.
The Thematic Analysis revealed four major themes
related to the human element and autonomy in com- shipping control configurations having to share the
plex safety-critical systems: (i) Trust; (ii) Awareness & same physical environment (i.e. traditionally manned
Understanding; (iii) Control, and (iv) Training and and operated ships, fully autonomous unmanned ships,
Organisation of Work. Several other themes emerged, unmanned manual shore-controlled ships, fishing ves-
although with less frequency and attention across sels, pleasure crafts, etc.). The participants highlighted
participant interviews, which are categorised within that the actors must be able to trust each other and
an additional theme: (v) Practical Implementation each other’s technological systems in order to operate
Considerations (see Figure 2). This fifth theme consists confidently and successfully.
of various sub-topics that arose where participants dis-
cussed pragmatic threats and limitations, not of the 4.2. Awareness and understanding
technology, or technological development itself, but It was revealed that the concept of ‘trust’ in autono-
rather of real-world implementation issues regarding mous systems was closely associated with the require-
such areas as regulations, cybersecurity and economic ment for a certain level of awareness and
considerations. understanding of a particular system, its operations
and what/how decisions are made. Several participants
4.1. Trust saw the development of trust as a function of aware-
ness and understanding of a system and its decision-
A predominant theme throughout participant inter- making. For example, in shipping, an awareness of
views was the concept of trust. Participants stressed external entities (i.e. the operations of other ships and
the need to trust the operations and decisions of external actors) is achieved through common rules
autonomous systems as human operators continue to and regulations (e.g. The IMO’s industry standard reg-
move further away from the sharp-end of processes. ulations all ships/operators must abide by [i.e. STCW,
Trust was a pervasive topic that transcended most COLREGS, etc.]). These common regulatory frameworks
areas of participant interviews, and the subsequent standardise procedures, and thus create shared
themes presented in the results. expectations and promotes trust between seafarers,
Trust was viewed as a function of successful and pro- ships and other stakeholders. Participants felt these
ven operations, which participants discussed in terms of qualities were necessary between humans and emerg-
societal norms and acceptance in technology. The devel- ing technologies, with one participant stating:
opment of trust in autonomous applications in a com-
‘The system needs to know your intentions. If you are
plex safety-critical domain, such as shipping, requires acting as a team member or team it is very important
the acceptance from many actors. For example, several that you know what the other persons intention is, so
participants discussed the concept of multi-modal you can support that and you can also predict what
ERGONOMICS 339

the other one needs … you need an idea of the Participants asked if the traditional seafaring career,
intention of the other system or the person.’ and thus if a traditional seafaring education would be
Participants discussed that regardless of whether relevant in the future. They proposed differing future
humans are in active control of a system, or in more skills that will be required highlighting the perceived
passive monitoring and supervisory roles, they require likelihood that programming and serious gaming skills
some level of understanding of how an autonomous will be relevant. One participant likened future seafar-
system works, its capabilities, strengths, weaknesses, ing to that of Air Traffic Controllers or Vessel Traffic
and ultimately knowledge of how decisions are made. Service operators where a highly specialised education
In summary, it was emphasised by the participants with differing skills than today would be required.
that the future operators working in these systems
should have the ability to ‘zoom out’ and think at
4.5. Practical implementation considerations
more abstract systemic level to enable efficient prob-
lem solving and troubleshooting if required. The fuzzier fifth theme encompassed several sub-
topics, which related to the real-world practical con-
siderations of implementing autonomous technologies
within the maritime domain. Several topics were men-
4.3. Control
tioned, including cybersecurity, economic, safety,
Participants questioned the value, relevance and pur- environmental and ethical issues. However, the pre-
pose of humans in increasingly automated and dominant sub-topic addressed the rules and regula-
autonomous systems of the future. Although shipping tions of international autonomous shipping:
operations, in general, have become increasingly auto-
‘The technology is already there, it is just the question
mated over the past several decades, with seafarers
of implementing it. The main issue is going to be
taking on increasingly supervisory roles, participants regulations, particularly in the maritime domain … ’
stressed the concept of whether the relationship
between humans and autonomous systems was one Participants were generally impressed with both the
of collaboration or one of support. In particular, sev- pace and direction of technological development, how-
eral participants noted that systems are generally reli- ever, continually brought up the conservative nature of
able and trustworthy in normal operations, however the international regulatory body and the possibility of
when abnormal situations occur, humans have many inhibiting real-world application within the industry. This
strengths, including the ability to adapt, be creative was a general reflection of the human perspective in
and think ‘outside of the box’. Furthermore, as autonomous systems that was discussed across partici-
humans take on increasingly supervisory roles one pants: are we moving too fast and do we know the
participant feared the loss of relevant skills and know- impacts of these emerging technologies? As one partici-
ledge of operations: pant made a cautionary remark of ‘unknown unknowns’,
‘(we need to) generally prevent them (operators) from
stressing the inherent lack of understanding at present
becoming deskilled. This is a huge problem with as how these systems may impact safety in daily opera-
automation because once the system breaks down tions that is difficult to conceptualise at present.
you are so untrained you have forgotten all about
how to actually work with it.’
5. Discussion
5.1. Future roles of the human element in the
4.4. Training and organization of work Maritime domain
‘There would be new kinds of jobs, there will be a
new kind of careers.’
The results from participant interviews revealed five
prevalent themes regarding the issues of implementing
Several of the participants (particularly those working autonomous systems and the future role of humans in
with training and assessment research and within pro- the maritime industry. With the incremental shift of
fessional training facilities) discussed the types of skills operations to increasingly automated and autonomous
people involved with autonomous systems will require functioning, future maritime operators are finding
in the future: themselves in progressively supervisory (and/or redun-
‘I think so far everyone is very occupied with the dant) roles that are physically separated from sharp-end
technology but not with how we train people in using operations. Traditional seafaring skills may become
these systems.’
unnecessary and obsolete, just as in the past differing
340 S. C. MALLAM ET AL.

maritime skills have evolved and changed with techno- required for successful operation in a multi-modal
logical advancements (Mallam and Lundh 2016). operating paradigm. Today, trust and understanding is
Participants discussed what type of profile and skillset a established between seafarers and ships with an
future ship operator may need. This ranged from a trad- underlying knowledge that operators are certified (e.g.
itional seafaring education, certification and at-sea hold valid certificates and minimum levels of compe-
experience, to non-seafarers who have a computer sci- tence) and are following a shared set of rules, similar
ence background with coding skills, to video game to the trust motorists put in each to operate within a
enthusiasts comfortable with command and control of shared space (i.e. the road, traffic signals, signage,
virtual agents and virtual worlds. etc.). In addition, there have been concerns voiced by
Future maritime operators may never go to sea researchers and practitioners regarding the need for
themselves, but instead, receive training of ship opera- cultivation of soft skills, such as creativity, leadership
tions remotely through simulator exercises or shore- and collaboration amongst the operators working in
side centres. Despite the argument that removing tandem with automated agents and Artificial
humans from ships will reduce ‘human error’, which is Intelligence in order to better leverage the traditional
attributed to the majority of maritime accidents, this strengths of humans in socio-technical systems
simply reorganises humans within the system, reduc- (Bainbridge 1983; Bradshaw et al. 2013; Deming 2017).
ing or eliminating some weaknesses but also creating Similar issues are faced across transport domains.
new and differing threats to safety (Wahlstro €m et al. For example, a feature of autonomous automobiles
2015). As these operators likely will obtain a supervis- that presents a unique challenge is the Take Over
ory role, the question of required skills is highly rele- Request; the event when the automated system
vant. The participants in the interviews revealed a expects the driver to take manual control of vehicle.
common prediction of skills related to programming, The driver, after a period of supervision, is notified to
Information Technology, analytics and serious gaming take control of the automobile within a certain time
being valuable assets. However, participants also limit. In this scenario, prior familiarisation with the sys-
described how basic knowledge and experience of tem and its limitations help users maintain trust and
seafaring would be needed to comprehend the set expectations (Beggiato and Krems 2013). This trust
actions of autonomous systems in order to interact and knowledge of the system is relevant for both the
and interpret operations when necessary. operators and passengers within the automobile, as
These interactions will also require humans to have a well as external agents in the surrounding area (e.g.
level of trust in highly automated and autonomous sys- fellow drivers, automobiles, pedestrians, cyclists shar-
tems (Lee and See 2004). According to Lee and ing the roadways). In the aviation sector, the type of
Moray (1992) there are four dimensions of trust: (1) foun- information provided to pilots affects their trust in air-
dation (2) performance (3) process and (4) purpose. craft systems. Ferris, Sarter, and Wickens (2010) found
Technological systems are typically predictable and reli- a benefit to providing pilots information related to the
able in order to satisfy foundation and purpose dimen- likelihood or confidence measures of an event, rather
sions. However, due to the use of sophisticated than a reporting a binary presence or absence of dan-
algorithms in adaptive automation systems, the response ger. This was found to better assist pilots in correctly
of a system (i.e. performance and processes) might not diagnosing and addressing abnormal situations. A
be what the user had anticipated (Inagaki 2003). This similar approach might help in supporting seafarers
phenomenon is referred to as an ‘automation surprise’ working with autonomous systems onboard ships and
(Sarter, Woods, and Billings 1997; Wickens 1995). help users build trust and knowledge in how autono-
Therefore, it will be necessary to support the trust of sea- mous systems make decisions and operate. Prior famil-
farers with the technological systems in dimensions of iarisation to the capabilities and limitations of a
performance and process with increasing automation. system will be necessary to instill trust, thus highlight-
Trust will become an increasingly important aspect ing the importance of developing new education and
in the future of surface activity at sea as differing enti- training programmes based on new and evolving skills
ties sharing the same physical space operate under required for ‘seafarers’ of the future.
different types of control. A shared waterway may be
occupied by an endless combination of manned and
5.2. Thinking outside the box
unmanned structures with varying levels of automa-
tion and autonomous technologies. Thus, a level of Highly automated ships have been predicted to operate
trust and understanding in external systems is with increased safety levels in comparison to
ERGONOMICS 341

contemporary manned ships, with the intention of to overcome, rather than technological aspects related
removing the human, and thus ‘human error’ to autonomy. This lies partially in the history of how the
(Hetherington, Flin, and Mearns 2006). Under normal IMO and other regulatory bodies have evolved to man-
operating parameters it is likely that ship operations age the international shipping industry. Maritime safety
will be generally successful. However, limitations may policy has traditionally been highly reactive, generally
be revealed if a ship encounters a situation completely developing in response to specific accidents or incidents
unfamiliar and unrelated to previous experiences or (Schro€der-Hinrichs et al. 2013). Regulatory bodies may
expectations, or where no official procedures are yet in also have conflicting relationships between balancing
place. As discussed by Lutzhoft and Dekker (2002), safety and economic interests of the stakeholders it rep-
automation can create new error pathways, prevent resents (Rosario 2000). A prescriptive regulatory regime
error detection and transfer its consequences. In such may not be suitable for autonomous ships, due to the
abnormal situations, an autonomous system may be pace of technological development, and would thus
deficient, whereas a human could employ creative and require more goal-based and functional requirements
adaptive solution-oriented thinking and decision-mak- (Mahapatra 2018). Furthermore, questions regarding the
ing. Humans may be relatively unpredictable agents in responsibility, liability and insurance of autonomous
complex systems, but they are also able to interpret sit- ships and shipping are still unclear (Mahapatra 2018).
uations and find solutions where current technologies The rapid developments in autonomous technologies
may falter. The predicted increase in safety may be and ongoing autonomous shipping projects is forcing
highly dependent on the flexibility and adaptability of the IMO and associated stakeholders to rethink its
the system. Future applications of Artificial Intelligence approach in dealing with technology in operations, and
and machine learning may negate any strengths and their far-reaching impacts on the industry as a whole.
advantages of having humans within complex opera- The push for autonomous shipping is being led by
tions with systems that are able to adapt, optimise and technologists while other maritime stakeholders, in par-
learn from vast amounts of data and scenarios. ticular, regulatory bodies and insurers, are struggling to
Just as traditional shipbuilding performed today, keep pace. Similarly, from a pedagogic and competency
the design and development of autonomous ships is development perspective maritime education and train-
implemented and controlled by humans. Thus, a ships’ ing institutions are confronted with the need to adapt
quality is dependent on the decision-making, compe- future courses and programmes for cadets and profes-
tencies and constraints of a particular shipbuilding sionals working within maritime industries. New training
project, ranging from initial planning, design and con- and education programmes will need to be developed
struction. Any ‘error’ or compromise (whether foreseen to match learning objectives with the new and emerg-
or unforeseen) introduced in the design of automated ing operational demands created by unmanned and
systems, therefore, contributes to potential unpredict- autonomous maritime systems.
ability and safety issues later in the ships’ lifecycle Utilising autonomous ships could save costs by
(Ahvenjaravi 2016). The occurrence of navigational reducing or eliminating crews (Rødseth and Burmeister
accidents (e.g. collisions and groundings) is expected 2012), including the facilities required for having
to decrease in unmanned ships, however, the rate of humans onboard (i.e. hotel facilities [kitchen, dining hall,
non-navigational accidents (e.g. onboard fires and sys- sleeping quarters, recreational areas, toilets and show-
tem failures) could increase due to the complexity of ers, etc.] through savings on weight, space and con-
the autonomous ships and their onboard equipment sumption of electricity, freshwater and supplies). Yet
(Wrobel, Montewka, and Kujala 2017). these ships will likely still require some level of supervi-
sion, as well as maintenance of onboard systems (if and
when equipment breaks down at sea what procedures
5.3. The way forward
would be in place?). If the reduction of onboard crew
The introduction of autonomous shipping requires leads to increased onshore manning, cost savings
amendments of many international treaties (e.g. SOLAS, related to crew reduction may be minimal, as the crew
STCW, IMDG, Polar Code, etc.), necessitating a departure would simply be ‘moved’ onshore. Furthermore, the
from well-established regulatory frameworks (IMO increased costs of specialised equipment, sensors,
2018a). This presents significant regulatory barriers to Information Technology infrastructure, transport to/from
the practical implementation of autonomous shipping remote ships at sea for unforeseen maintenance,
(DMA 2017). The participants discussed this practical amongst other issues, may also negate any cost savings
implementation issue, feeling that it is the largest barrier argued for by reducing or eliminating crew.
342 S. C. MALLAM ET AL.

Autonomy is a highly disruptive technology for an the relatively narrow research scope and questions the
industry that is generally considered conservative and sample provided rich descriptions from which thematic
traditional. Autonomous technologies have created saturation occurred (Saunders et al. 2018) and clearly
emerging operational paradigms that are creating defined themes emerged from the data analysis (Braun
completely new possibilities for how to work within and Clarke 2006). The data was collected and analysed
maritime transport is organised and performed. The using the six-step Thematic Analyses framework (Braun
maritime domain must also look across disciplines to and Clarke 2006). All interviews were audio recorded
the research and development performed in other sec- and transcribed verbatim. Three researchers independ-
tors of transport and industries who have greater ently analysed the transcripts and then convened in an
experience with both developing and applying interactive workshop to validate their individual find-
autonomous systems in the real-world. This can better ings and finalise the results. Having multiple coders
enable the successful implementation of autonomous work independently increases the validity and reliabil-
technologies to the maritime domain. ity of the data analyses through inter-coder agreement
(Corbin and Strauss 2008; MacQueen et al. 1998).
5.4. Methodological discussion and limitations However, ensuring sufficient validity and reliability in
qualitative interview data is challenging and qualitative
The qualitative research design utilised a semi-
interviewing has its limitations as a research instrument
structured interview format for individual participant
(Alshenqeeti 2014). In particularly, the external validity
interviews. This provided an opportunity to elicit descrip-
of our research findings may be difficult to generalise
tive responses to questions within our broadly defined
to other domains, as autonomy and autonomous tech-
scope of three areas of interest: (1) autonomy definition
nologies and systems are rapidly developing and much
and concepts, (2) perceived challenges and changes
of the discussion was focussed on maritime-specific
with autonomous systems, and (3) the role of humans
issues. Regardless, at this early stage in both research
within autonomous systems. The semi-structured inter-
and real-world application for autonomous maritime
views allowed for more exploratory and descriptive
applications, this exploratory study provides initial
answers from participants (Corbin and Strauss 2008).
input and frameworks for areas of interest regarding
Furthermore, in-person interviews offered the opportun-
the human element and autonomous maritime sys-
ity for interviewees to clarify or rephrase questions in
order to achieve mutual understanding, and thus more tems and operations.
accurate data (Do €rnyei 2007). As the sample was highly
international and interviews were carried out in both 6. Conclusions
English and Norwegian (for participants who were native
The pace of the technological development of autono-
speakers), the semi-structured interview format allowed
mous systems is rapid, resulting in current scaled pilot
for immediate clarification between researcher and
interviewee. projects and full-scale autonomous ships under devel-
The participant sample consisted of ten SME’s work- opment. The obstacles of implementing such solutions
ing within academia (five) and as industry practitioners are no longer necessarily technological, but rather
(five). The inherent individual characteristics and back- related to human-technology aspects of new operating
ground of the recruited participants, as well as the paradigms and emerging regulatory, liability and secur-
interview manuscript scope and interview sessions ity concerns. Despite recent attention and practical
may have primed participants and biased (e.g. expect- demonstrations of utilising autonomous technologies
ancy bias) their responses towards more human fac- across other transport domains, challenges continue to
tors-centric topics than typical. The educational and reveal themselves. Issues relating to system trust,
professional profiles of the participants varied widely, understanding and predictability of decision-making, as
but all ten had strong connections to human element well as the skills required to develop, operate and
issues in highly automated and autonomous safety- maintain such technologies will become increasingly
critical systems within their current work. relevant to the successful and sustainable functioning
Due to the specialisation of the research and the of highly automated and autonomous maritime sys-
desire to have SME participants who are actively tems. As autonomous technologies continue to rapidly
engaged in safety-critical domains, recruitment for the advance, it is critical to develop a better understanding
data collection was limited to ten. Although, this num- of how and where humans fit into operations in ever
ber of participants may be considered low, because of evolving complex safety-critical systems.
ERGONOMICS 343

Acknowledgement Braun, V., and V. Clarke. 2006. “Using Thematic Analysis in


Psychology.” Qualitative Research in Psychology 3(2):
This research was partially supported by MARKOM2020 77–101. doi:10.1191/1478088706qp063oa.
under project number 1700111 (2017). The authors would Chauvin, C., S. Lardjan, G. Morel, J. P. Clostermann, and B.
like to thank Sunniva Veie for her valuable contributions in Langard. 2013. “Human and Organizational Factors in
assisting with data collection and transcription. This paper Maritime Accidents: Analysis of Collisions at Sea Using the
has been expanded from an earlier version presented at the HFACS.”Accident Analysis and Prevention 59: 26–37. doi:10.
20th Congress of the International Ergonomics Association 1016/j.aap.2013.05.006.
in Florence, Italy (August 26th–30th, 2018) and published Corbin, J., and A. Strauss. 2008. Basics of qualitative research.
within its conference proceedings: 3rd ed. Thousand Oaks, California: Sage Publications.
Mallam S. C., Nazir S., Sharma A. & Veie S. (2019). Deming, D. J. 2017. “The Growing Importance of Social Skills
Perspectives on Autonomy – Exploring Future Applications in the Labor Market.” The Quarterly Journal of Economics
and Implications for Safety Critical Domains. In: Bagnara S., 132(4): 1593–1640. doi:10.3386/w21473.
Tartaglia R., Albolino S., Alexander T., Fujita Y. (eds). DMA (Danish Maritime Authority). 2017. Analysis of
Proceedings of the 20th Congress of the International Regulatory Barriers to the Use of Autonomous Ships: Final
Ergonomics Association (IEA 2018). IEA 2018. Advances in Report. Copenhagen, Denmark: Danish Maritime Authority.
Intelligent Systems and Computing, vol 82, pp. 396–405. Doraszelski, U. 2004. “Innovations, Improvements, and the
Springer, Cham. DOI: 10.1007/978-3-319-96080-7_47 Optimal Adoption of New Technologies.” Journal of
Economic Dynamics and Control 28(7): 1461–1480. doi:10.
1016/S0165-1889(03)00112-X.
Disclosure statement Do€rnyei, Z. 2007. Research Methods in Applied Linguistics.
No potential conflict of interest was reported by the authors. Oxford, United Kingdom: Oxford University Press.
Endsley, M. R., and E. O. Kiris. 1995. “The out-of-the-Loop
Performance Problem and Level of Control in
References Automation.” Human Factors: The Journal of the Human
Factors and Ergonomics Society 37(2): 381–394. doi:10.
Ahvenjaravi, S. 2016. “The Human Element and Autonomous 1518/001872095779064555.
Ships.” TransNav 10(3): 517–523. doi:10.12716/1001.10.03.18. Ferris, T., N. Sarter, and C. D. Wickens. 2010. “Cockpit
Allen, P. 2009. “Perception of Technology at Sea Amongst Automation: Still Struggling to Catch Up.” 2nd ed. In E.
British Seafaring Officers.” Ergonomics 52(10): 1206–1214. Salas, T. Allard, D. Maurino (Eds.), Human factors in
doi:10.1080/00140130902971924. Aviation, 479–503. Burlington, MA, USA: Academic Press.
Allianz (Allianz Global Corporate & Specialty). 2012. Safety doi:10.1016/B978-0-12-374518-7.00015-8.
and Shipping 1912–2012: From Titanic to Costa Concordia. Grech, M., T. J. Horberry, and T. Koester. 2008. Human
Munich, Germany: Allianz Global Corporate & Specialty. Factors in the Maritime Domain. Boca Raton, FL, USA: CRC
Allianz (Allianz Global Corporate & Specialty). 2017. Safety Press.
and Shipping Review 2017. Munich, Germany: Allianz Hetherington, C., R. Flin, and K. Mearns. 2006. “Safety in
Global Corporate & Specialty. Shipping: The Human Element.” Journal of Safety Research
Allianz (Allianz Global Corporate & Specialty). 2018. Safety 37(4): 401–411. doi:10.1016/j.jsr.2006.04.007.
and Shipping Review 2018. Munich, Germany: Allianz Hoff, K. A., and M. Bashir. 2015. “Trust in Automation:
Global Corporate & Specialty. Integrating Empirical Evidence on Factors That Influence
Alshenqeeti, H. 2014. “Interviewing as a Data Collection Trust.” Human Factors: The Journal of the Human Factors
Method: A Critical Review.” English Linguistics Research and Ergonomics Society 57(3): 407–434. doi:10.1177/
3(1): 39–45. doi:10.5430/elr.v3n1p39. 0018720814547570.
Bainbridge, L. 1983. “Ironies of Automation.” Automatica Hummels, D. 2007. “Transportation Costs and International
19(6): 775–779. doi:10.1016/0005-1098(83)90046-8. Trade in the Second Era of Globalization.” Journal of Economic
Beggiato, M., and J. F. Krems. 2013. “The Evolution of Mental Perspectives 21(3): 131–154. doi:10.1257/jep.21.3.131.
IMO (International Maritime Organization). 2018a. Report of
Model, Trust and Acceptance of Adaptive Cruise Control
the Maritime Safety Committee on its Ninety-Ninth Session.
in Relation to Initial Information.”Transportation Research
London, United Kingdom: International Maritime
Part F: Traffic Psychology and Behaviour 18: 47–57. doi:10.
Organization.
1016/j.trf.2012.12.006. IMO (International Maritime Organization) 2018b. “IMO takes
Bonner, M., R. Taylor, K. Fletcher, and C. Miller. 2000.
first steps to address autonomous ships.” Accessed
“Adaptive Automation and Decision Aiding in the Military September 18 2018. http://www.imo.org/en/mediacentre/
Fast Jet Domain.” In Proceedings of the Conference on pressbriefings/pages/08-msc-99-mass-scoping.aspx
Human Performance, Situation Awareness and Inagaki, T. 2003. “Adaptive Automation: Sharing and Trading
Automation: User-Centered Design for the New of Control.” In E. Hollnagel (Ed.), Handbook of cognitive
Millennium, 154–159. task design, 147–169. Mahwah, NJ, USN: Lawrence
Bradshaw, J. M., R. R. Hoffman, M. Johnson, and D. D. Erlbaum Associates, Publishers.
Woods. 2013. “The Seven Deadly Myths of ‘Autonomous Kim, T. E., S. Nazir, and K. I. Øvergård. 2016. “A STAMP-Based
Systems.” IEEE Intelligent Systems 28(3): 54–61. doi:10. Causal Analysis of the Korean Sewol Ferry Accident.”
1109/MIS.2013.70. Safety Science 83: 93–101. doi:10.1016/j.ssci.2015.11.014.
344 S. C. MALLAM ET AL.

KM (Kongsberg Maritime) 2018. “Autonomous ship project, Automation.” IEEE Transactions on Systems, Man, and
key facts about YARA Birkeland.” Accessed September 13 Cybernetics – Part A: Systems and Humans 30(3): 286–297.
2018. https://www.km.kongsberg.com/ks/web/nokbg0240. doi:10.1109/3468.844354.
nsf/AllWeb/4B8113B707A50A4FC125811D00407045?Open Progoulaki, M., and M. Roe. 2011. “Dealing with Multicultural
Document Human Resources in a Socially Responsible Manner:
Kyriakidis, M., J. C. G. de Winter, N. Stanton, T. Bellet, B. van A Focus on the Maritime Industry.” WMU Journal
Arem, K. Brookhuis, M. H. Martens, K. Bengler, J. Andersson, of Maritime Affairs 10(1): 7–23. doi:10.1007/s13437-011-
N. Merat., N. Reed, M. Flament, M. Hagenzieker, and R. 0003-0.
Happee. 2019. “A Human Factors Perspective on Automated Richards, D., and A. Stedmon. 2016. “To Delegate or Not to
Driving.” Theoretical Issues in Ergonomics Science 20(3): Delegate: A Review of Control Frameworks for Autonomous
223–249. doi:10.1080/1463922X.2017.1293187. Cars.” Applied Ergonomics 53: 383–388. doi:10.1016/j.apergo.
Lee, J. D., and N. Moray. 1992. “Trust, Control Strategies and 2015.10.011.
Allocation of Function in Human Machine Systems.” Rosario, J. L. 2000. The Role of Maritime Administration in
Ergonomics 35(10): 1243–1270. doi:10.1080/00140139 Ensuring Quality Shipping Services: A Guide to Cape Verde
208967392. Maritime Administration. Malmo €, Sweden: World Maritime
Lee, J. D., and K. A. See. 2004. “Trust in Automation: University.
Designing for Appropriate Reliance.” Human Factors: The Rødseth, Ø. J., and H. C. Burmeister. 2012. “Developments
Journal of the Human Factors and Ergonomics Society Towards the Unmanned Ship.” Paper presented at the
46(1): 50–80. doi:10.1518/hfes.46.1.50_30392. International Symposium “Information on Ships” (ISIS),
Lutzhoft, M. H., and S. W. Dekker. 2002. “On Your Watch: Hamburg, Germany, August 30–31.
Automation on the Bridge.” Journal of Navigation 55(1): Rødseth, O. J., and H. C. Burmeister. 2015. “Risk Assessment
83–96. doi:10.1017/S0373463301001588. for an Unmanned Merchant Ship.” Transnav, The
MacQueen, K. M., E. McLellan, K. Kay, and B. Milstein. 1998. International Journal on Marine Navigation and Safety of Sea
“Codebook Development for Team-Based Qualitative Transportation 9(3): 357–364. doi:10.12716/1001.09.03.08.
Analysis.” Cam Journal 10(2): 31–36. doi:10.1177/ Sarter, N. B., D. D. Woods, and C. E. Billings. 1997.
1525822X980100020301. “Automation Surprises.” 2nd ed. In G. Salvendy (Ed.),
Mahapatra, A. 2018. “Digitalization and Maritime Autonomous Handbook of Human Factors and Ergonomics. 1926–1943.
Surface Ships (MASS).” Accessed September 29 2018. New York, USA: Wiley.
https://www.dma.dk/Vaekst/autonomeskibe/Documents/ Saunders, B., J. Sim, T. Kingstone, S. Baker, J. Waterfield, B.
Ashok%20Mahapatra%20Maritime%20Autonomous%20Surface Bartlam, H. Burroughs, and C. Jinks. 2018. “Saturation in
%20Ships.pdf Qualitative Research: Exploring Its Conceptualization and
Mallam, S. C. 2016. Distributed participatory design in multidis- Operationalization.” Quality and Quantity 52(4):
ciplinary engineering projects: Investigating a sustainable 1893–1907. doi:10.1007/s11135-017-0574-8.
approach for ship design and construction. Gothenburg, Schro€ der-Hinrichs, J. U., E. Hollnagel, M. Baldauf, S. Hofmann,
Sweden: Chalmers University of Technology. and A. Kataria. 2013. “Maritime Human Factors and IMO
Mallam, S. C., and M. Lundh. 2016. “The Physical Work Policy.” Maritime Policy and Management 40(3): 243–260.
Environment and End-User Requirements: Investigating doi:10.1080/03088839.2013.782974.
Marine Engineering Officers Operational Demands and Ship Sheridan, T. B. 1992. Telerobotics, Automation, and Human
Design.” Work 54(4): 989–1000. doi:10.3233/WOR-162365. Supervisory Control. Cambridge, MA, USA: MIT Press.
Mallam, S. C., M. Lundh, and S. N. MacKinnon. 2017. Society of Automotive Engineers. 2018. “Taxonomy and
“Evaluating a Digital Ship Design Tool Prototype: Designers’ Definitions for Terms Related to Driving Automation
Perceptions of Novel Ergonomics Software.” Applied Systems for On-Road Motor Vehicles” (J3016_201806).
Ergonomics 59: 19–26. doi:10.1016/j.apergo.2016.08.026. Accessed June 23 2019. https://www.sae.org/standards/
Nazir, S., A. Kluge, and D. Manca. 2014. “Automation in content/j3016_201806/
Process Industry: Cure or Curse? How Can Training Stopford, M. 2009. Maritime Economics. 3rd ed. Oxon, United
Improve Operator’s Performance.” Computer Aided Kingdom: Routledge.
Chemical Engineering 33: 889–894. doi:10.1016/B978-0- Strauch, B. 2017. “The Automation-by-Expertise-by-Training
444-63456-6.50149-6. Interaction: Why Automation-Related Accidents Continue
NFAS (Norwegian Forum for Autonomous Ship). 2017. to Occur in Sociotechnical Systems.” Human Factors: The
Definitions for Autonomous Merchant Ships. Trondheim, Journal of the Human Factors and Ergonomics Society
Norway: Norwegian Forum for Autonomous Ships. 59(2): 204–228. doi:10.1177/0018720816665459.
Parasuraman, R. 1987. “Human-Computer Monitoring.” Human Wahlstro €m, M., J. Hakulinen, H. Karvonen, and L. Lindborg.
Factors: The Journal of the Human Factors and Ergonomics 2015. “Human Factors Challenges in Unmanned Ship
Society 29(6): 695–706. doi:10.1177/001872088702900609. Operations – Insights from Other Domains.” Procedia
Parasuraman, R., and V. Riley. 1997. “Humans and Manufacturing 3: 1038–1045. doi:10.1016/j.promfg.2015.07.167.
Automation: Use, Misuse, Disuse, Abuse.” Human Factors: Wickens, C. D. 1995. “Designing for Situation Awareness and
The Journal of the Human Factors and Ergonomics Society Trust in Automation.” IFAC Proceedings Volumes 28(23):
39(2): 230–253. doi:10.1518/001872097778543886. 365–370. doi:10.1016/S1474-6670(17)46646-8.
Parasuraman, R., T. B. Sheridan, and C. D. Wickens. 2000. “A Wilhelmsen , 2018. “Wilhelmsen and KONGSBERG Establish
Model for Types and Levels of Human Interaction with World’s First Autonomous Shipping Company”.
ERGONOMICS 345

Accessed October 14 2018. https://www.wilhelmsen. bel, K., J. Montewka, and P. Kujala. 2017. “Towards the
Wro
com/media-news-and-events/press-releases/2018/wilhelm- Assessment of Potential Impact of Unmanned Vessels on
sen-and-kongsberg-establish-worlds-first-autonomous-ship- Maritime Transportation Safety.” Reliability Engineering and
ping-company/ System Safety 165: 155–169. doi:10.1016/j.ress.2017.03.029.

You might also like