International Journal of Computer Integrated

Manufacturing

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tcim20

Industry 4.0: survey from a system integration

perspective

Manuel Sanchez, Ernesto Exposito & Jose Aguilar

To cite this article: Manuel Sanchez, Ernesto Exposito & Jose Aguilar (2020) Industry 4.0: survey
from a system integration perspective, International Journal of Computer Integrated Manufacturing,
33:10-11, 1017-1041, DOI: 10.1080/0951192X.2020.1775295

To link to this article: https://doi.org/10.1080/0951192X.2020.1775295

Published online: 25 Jun 2020.

Submit your article to this journal

Article views: 1524

View related articles

View Crossmark data

Citing articles: 33 View citing articles

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

https://www.tandfonline.com/action/journalInformation?journalCode=tcim20
INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING
2020, VOL. 33, NOS. 10–11, 1017–1041
https://doi.org/10.1080/0951192X.2020.1775295

ARTICLE

Industry 4.0: survey from a system integration perspective

a,b,c b c,d
Manuel Sanchez , Ernesto Exposito and Jose Aguilar
a
Informatics Engineering, Universidad Nacional Experimental del Táchira, San Cristóbal, Venezuela; bE2S UPPA, LIUPPA, Universite de Pau et
des Pays de l’Adour, Anglet, France; cCEMISID, Universidad de Los Andes, Mérida, Venezuela; dGIDITIC, Universidad EAFIT, Medellín, Colombia

ABSTRACT ARTICLE HISTORY

In recent years, a revolution named Industry 4.0 has arisen. Industry 4.0 is presented as the Received 27 December 2018
integration of new advances in areas such as Cyber-Physical Systems, the Internet of Things and Accepted 17 May 2020
Everything (IoE), Cloud computing, the Internet of Services, Big Data Analysis, Smart Factories, KEYWORDS
Augmented Reality, among others. Industry 4.0 is not only a new industrial revolution, but also Industry 4.0; integration;
a crucial integration challenge that involves several actors from the IoE, which are people, data, interoperability; IoE; 5C
services, and things. This paper proposes an approach to analyze the integration challenges in the
context of Industry 4.0 using five integration levels, which are connection, communication,
coordination, cooperation, and collaboration (5 C). In that sense, this paper presents a state of
the art of recent studies in Industry 4.0 from an integration perspective, categorized according to
the 5 C integration levels versus the four actors of IoE. Specifically, this paper considers several
works intended to solve problems of autonomic integration in Industry 4.0 at the highest levels of
the 5 C integration stack (coordination, cooperation, and collaboration). Also, this paper presents
a case study from an integration perspective, which contemplates autonomy, self-organizing,
among other aspects, in order to turn a traditional industry into a smart factory regarding the
Industry 4.0 concept.

1. Introduction ● The Internet of Things (IoT). It is a concept that

refers to the connections between physical
Because of the digital revolution, the boundaries
objects, such as sensors or machines, and the
between the physical and digital worlds are shrinking
Internet (Sengupta, Gupta, and Vinayak 2017;
to give life to a more interconnected and smart factories
Riggins and Keskin 2017).
in which employees, machines, processes, and products,
● Internet of Everything (IoE). It is an evolution of
interact to give a better organization of all the produc­
IoT that refer to connect not only things, but also
tive means, to empower the entire company itself to
people, processes, and data, all connected to the
achieve higher levels of efficiency and productivity. In
Internet (D. Lee, Choi, and Kim 2017).
that sense, the concept of Industry 4.0, is used to desig­
● Human-Machine Interaction (HMI). It defines the
nate the new generation of connected, robotics, and
interfaces that allow humans to take part in
intelligent factories. Fundamentally, the vision of
a system but removing the risk that operations
Industry 4.0 is to give smart capabilities to the produc­
can represent for their lives.
tion and physical operations in order to create a more
● X-Mining (Everything mining). It refers to any
holistic and better-connected ecosystem. Figure 1 pre­
data analytics technique that could be applied
sent the technologies typically used to bring solutions in
to the system actors (people, things, processes
Industry 4.0, such as (see section 2.2 for details):
and data), in order to get a better understanding
of the system. Techniques as data mining,
● System Integration: It refers to link together sys­
semantic mining, ontological mining, process
tem components (vertical integration), two or
mining, services mining, social mining, big data
more systems (horizontal integration), or to pro­
analytics, machine learning, among others, enter
vide interfaces to link physical and virtual objects
under this concept.
of a system (end to end integration). See (Suri
● Smart factory. It is a factory that invests and
et al. 2017; Pisching et al. 2018) for details
benefits from the technologies, solutions, and

CONTACT Manuel Sanchez mbsanchez@unet.edu.ve Informatics Engineering Universidad Nacional Experimental Del Táchira, San Cristóbal, Venezuela
© 2020 Informa UK Limited, trading as Taylor & Francis Group
1018 M. SANCHEZ ET AL.

Figure 1. Typical technologies in the Industry 4.0.

approaches of Industry 4.0 (Liu et al. 2017). Smart factories in shorter cycles than the traditional ones,
Factories are developed into intelligent environ­ making an efficient use of human and physical
ments in which the real and digital are fully- resources (Suri et al. 2017) while increasing the effec­
interconnected (Weyer et al. 2015). tive communication between the actors involved in
● Cloud computing. It consists of using intercon­ the production process (X. Li et al. 2017a). Particularly,
nected remote servers hosted on the Internet to Figure 1 shows that Industry 4.0 is a concept that
store, manage, and process information, using allows the integration of a vast diversity of Artificial
a service-oriented architecture (SOA) (Shila et al. Intelligence (AI) techniques and automation technol­
2017; Vaquero et al. 2008). ogies within the manufacturing domain to make
● Cyber-physical systems (CPS). It consists of the smarter organizations (Suri et al. 2017; X. Li et al.
characterization of each physical object using 2017a; Khan et al. 2017; Gökalp, Şener, and Eren 2017).
a cyber component, so that this last can act as Essentially, as Industry 4.0 is an emerging concept,
the smart part, in order to make decisions, inter­ there are a multitude of challenges, risks and barriers
operate, and execute actions in representation of limiting its implementation that need to be solved
the physical object associated with it (Goossens (Hofmann and Marco 2017; Schwab 2016; Lu 2017;
and Richard 2017). Preuveneers and Ilie-Zudor 2017; X. Li et al. 2017a;
Liao et al. 2017; Suri et al. 2017). In that sense, in
In general, form Figure 1 can be deduced that solu­ Figure 2, it can be seen the most common challenges
tions in Industry 4.0 use commonly Data mining tech­ around Industry 4.0., such as the complexity of the
niques, Big Data Analytics, Social mining, Service planning, for which it must be created explanatory
mining, among other mining techniques to build models for managing complex products and produc­
knowledge-bases oriented to make autonomous and tion systems (Liao et al. 2017). Another challenge
smarter decisions. Particularly, the application of related to Industry 4.0 corresponds to standardiza­
these technologies in Industry 4.0 allows the manage­ tion, which means to develop common standards to
ment of the production process autonomously and to support 5 C processes of connection, communication,
increase the factory productivity, flexibility, adaptabil­ coordination, cooperation, and collaboration (Liao
ity, and efficiency (X. Li et al. 2017a; Santos et al. 2017). et al. 2017). Another challenge is related to privacy
Moreover, the Industry 4.0 enables to reconfigure and security of data, due that sensors and smart
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1019

Figure 2. Common challenges in industry 4.0.

devices are continuously collecting information from supposes another challenge within the Industry 4.0
the environment, and it is necessary to protect that due to the heterogeneity of entities that generate
information and avoid unauthorized users to access a large amount of heterogeneous data that is not
that data, or still worse, to gain access to control the easy to homogenize, which consist of producing rele­
system (Preuveneers and Ilie-Zudor 2017; X. Li et al. vant information (especially in real-time). Another sig­
2017b; Liao et al. 2017). Additionally, the heterogene­ nificant challenge, related to integrability and
ity of data and actors is considered a big challenge (Lu interoperability issues, is how to provide autonomy
2017; Liao et al. 2017) in Industry 4.0, due that devices to the production process, in order to accomplish
from different manufacturers can generate data in global production goals. That means that entities
a diversity of formats and might communicate using involved in the production processes should be able
protocols that are incompatible with other brands. to make emerging autonomous coordination, coop­
Notably, in this paper, the integration and interoper­ eration, or collaboration processes (processes that
ability challenges throughout the 5 C stack levels are have not been explicitly specified). In this context,
considered. Integration is related to link together sys­ a central challenge is how to discover the coordina­
tem’s components or sub-systems to allow them to tion, cooperation, and collaboration necessities, and
act as a whole and unique system (Drăgan, Selea, and how to create a management plan to deploy these
Teodor-Florin 2017). processes, to allow autonomous integration pro­
On the other hand, interoperability is the ability of cesses at the 5 C highest levels for manufacturing.
two systems to understand each other and using In general, the authors believe that the integration
functionalities of one into another (Lu 2017), in such and interoperability challenges can be analyzed
a way that they can work together to produce useful according to the 5 C integration stack so that they
results adjusted to the integration goals. In other can be incrementally solved. For instance, issues
words, interoperability allows exchanging informa­ related to how to connect heterogeneous actors
tion between devices, business processes, interfaces, should be studied at the connection level, but issues
people, among others. (Santos et al. 2017; Lu 2017; related to the autonomous interoperability of actors
Khan et al. 2017), in order to solve conflicts and might be studied at the coordination, cooperation, and
achieve agreements in the execution of their tasks. collaboration levels, depending on the actors’ interac­
Fundamentally, integrability and interoperability tion. In that sense, this paper presents a survey regard­
requires from the entities involved to be able to con­ ing integrability and interoperability of actors in the
nect (to join each other), to communicate (to Industry 4.0 context. Besides, this paper discusses the
exchange information between each other), to coor­ challenges that might be solved in order to achieve
dinate (to follow the orders of a central entity in order integration and interoperability of actors, and finally, it
to achieve a global goal), to cooperate (to work with discusses the technologies that are commonly used in
others to achieve individual goals) and to collaborate order to solve those challenges. Consequently, the
(to work with others to achieve common goals). That research works discussed in the state of the art section
1020 M. SANCHEZ ET AL.

of this paper, are categorized according to the 5 C Subsequently, a second remarkable transformation
integration level that most fit them. It was made in of the industry, known as the second industrial revo­
that way because our goal is related to solving integra­ lution, started. This transformation process began at
tion and interoperability issues incrementally using the the end of the century XIX until the beginning of the
integration 5 C stack. century XX (Santos et al. 2017; X. Li et al. 2017a;
This paper is organized as follows. Section II pre­ Molano et al. 2017). According to (Santos et al. 2017;
sents as the first point, the evolution of the industry X. Li et al. 2017a; Molano et al. 2017; Huber and Weiss
towards the Industry 4.0, as well as the concepts and 2017), the elements that drove the second industrial
technologies related to it. Likewise, an evolution of revolution were the electricity and the division of
the IoE definition initially proposed by Cisco is pre­ labors. Those elements allowed manufacturing pro­
sented in order to allow expanding and generalizing ducts through assembly lines. Moreover, this revolu­
this paradigm. Subsequently, in section III, a short tion drove many changes, like the invention of the
state of the art focusing mainly on the integration internal combustion engine, the discovery of new
and interoperability of the actors of IoE within the sources of energy (electricity, oil, gas, among others.),
Industry 4.0 versus the 5 C levels (Connection, the telegraph, and the airplane, among others.
Communication, Coordination, Cooperation, and Industry 3.0 started in the 1960 s and extended
Collaboration), is presented. Moreover, in this section until the beginning of the century XXI (Santos et al.
is introduced a definition of these 5 C levels from our 2017; Molano et al. 2017). This revolution is also
point of view. In Section IV, a discussion of the chal­ known as the Digital Revolution (Santos et al. 2017)
lenges in Industry 4.0 versus the 5 C, is presented, because it was centered on the use of the information
finishing with some conclusions in Section V. technology (IT), electronic circuits and the Internet to
improve the production processes (Santos et al. 2017;
X. Li et al. 2017a; Huber and Weiss 2017). In this
2. Background
revolution, the Programmable Logic Circuits (PLC)
In this section is given a brief review of the evolution made possible, together with the Industrial
of the industry towards the fourth industrial revolu­ Automation (basically, directed by the AI), the integra­
tion, and introduces several fundamental concepts in tion of automatic machines into the production lines,
this context. allowing to reduce human errors and to increase the
development of products and services dramatically.
Likewise, Industry 3.0 allowed the creation of more
2.1. Introduction to industry 4.0
efficient, safer, and less polluting means of transport,
In ancient times of history, all aspects related to pro­ besides, the use of renewable energy was expanded,
duction, such as agriculture, transport, or the textile and the use of intelligent objects begun.
area, among others, were carried out manually, and In recent years, the Industry 4.0 or Fourth Industrial
the development of a product took a considerable revolution term was announced at the Hanover Trade
amount of time. An essential fact in the history that Fair, in 2011 (Santos et al. 2017; Romero et al. 2017).
had improved this situation considerably was the However, some authors affirm that the Fourth
emergence of the first industrial revolution. Industrial Revolution started at the beginning of the
The first industrial revolution started from century XXI (Santos et al. 2017; Molano et al. 2017).
the second half of the century XVIII and took until the Santos et al. (2017) affirm that this revolution is char­
mid of century XIX. It was driven by the creation of the acterized by the integration of artificial intelligence
water pump and the steam engine, which allowed to solutions within the production machines. Moreover,
mechanize the production and to transform manual X. Li et al. (2017a) affirm that the primary objective of
production processes into manufacturing processes Industry 4.0 is achieving high levels of operational
(Santos et al. 2017; X. Li et al. 2017a; Molano et al. efficiency and productivity. In the Industry 4.0 con­
2017; Huber and Weiss 2017). In such a way, the man­ text, new technologies, such as Cyber-Physical
ufacturing process allowed to transform raw materials Systems (CPS), Internet of Everything (IoE), Cloud
into products with the help of steam machines, redu­ computing, Augmented Reality, Big Data Analysis,
cing production times considerably. among others, are expected to play a crucial role in
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1021

order to enable factories to self-organize and self- ● Optimized: the production time is reduced, and the
control, in a distributed way and in real-time. use of human and physical resources is optimized.
Notably, that conjunction of technologies has allowed ● Transparent: new tools are used to support quick,
the creation of Smart Factories (Molano et al. 2017; transparent, and consistent decision-making pro­
Romero et al. 2017; Riel and Flatscher 2017), which cesses, as well as to track the orders appropriately.
can autonomously create smart products through ● Proactive: early identification of quality issues
smart processes and procedures (Molano et al. 2017). and anomalies allows self-planning and resche­
duling in real-time.
● Agile: adaptable layouts and equipment.
2.2. Conceptual framework of the industry 4.0

This section presents some fundamental concepts 2.2.2. Cyber-Physical Systems (CPS)
involved in Industry 4.0. CPS has been defined some time ago. This section
considers some recent definitions about this concept.
2.2.1. Smart factory CPS was defined by Zanero (2017) as a set of computa­
A Smart Factory defines innovative production mechan­ tional and physical interconnected resources, which
isms, which are suitable for a diversity of applications in uses a smart control loop in order to adapt and improve
different industrial branches involving modern produc­ the autonomy and efficiency of the whole system. In
tion technologies (Liu et al. 2017). Strozzi et al. (2017) the same sense, Qiu et al. (2017) affirm that CPSs are
affirm that factories become smarter, more efficient, characterized by strong interactions among cyber com­
safer, and more environmentally sustainable, thanks to ponents (software) and physical components (devices).
the intelligent combination and integration of produc­ Both elements combine artificial intelligence, automa­
tion technologies and devices, information and commu­ tion, and communication technologies tightly, in order
nication systems, data and services, and network to achieve high levels of performance, reliability, effi­
infrastructures. Also, Syberfeldt, Danielsson, and ciency, and robustness in a physical environment
Gustavsson (2017) say that the smart factory concept is (Goossens and Richard 2017). Likewise, Elattar, Wendt,
intended to enable extremely flexible, and self- and Jasperneite (2017) define CPS using a service-
adaptable production processes, with machines and oriented vision: ‘a CPS consists of one or more inter­
products that act both intelligently and autonomously, connected components or units, where services of each
by implementing concepts such as IoT and CPS. unit are visible to the other units of the system and
Moreover, Deloitte Consulting (2017) defines the smart allow them to cooperate.’ Jazdi (2014) affirms that
factory as one of the main features of Industry 4.0, which a CPS connected to the cloud is often referred to as
focuses on integrating various industrial devices to the ‘Internet of Things.’ On the other hand, a CPS is
establish a networked manufacturing system. Those essential in the context of Industry 4.0, because it helps
industrial devices interoperate autonomously in order to achieve production goals of the manufacturing pro­
to achieve the manufacturing goals. Mainly, a Smart cesses while improving the effectiveness and efficiency
Factory is a fully integrated industry (Romero et al. of the entire industry autonomously.
2017) that combines many technologies, such as 3D
Printing, AR, Radio-frequency identification (RFID), 2.2.3. Cloud computing and the Internet of Services
Enterprise Resource Planning (ERP), IoT, Smart predictive (IoS)
decision support tools, etc. (X. Li et al. 2017a; Strozzi et al. Cloud computing has emerged as a new computing
2017), which use the AI to increase the productivity and paradigm that offers excellent potential to share net­
efficiency of factories (Molano et al. 2017). Smart fac­ worked computing resources. Shila et al. (2017) define
tories combine several smart devices that are coordi­ Cloud computing as a revolutionary paradigm to deli­
nated, collaborate, and cooperate autonomously in ver computing resources, ranging from data storage/
order to achieve the production objectives. The main processing to software, as a service over the network,
features of a smart factory are (Deloitte Consulting 2017): typically using Internet technologies. Additionally,
(Shila et al. 2017; Vaquero et al. 2008) present the US
● Connected: The actors involved in the produc­ National Institute of Standards and Technology (NIST)
tion process can connect and communicate. referential architecture that has categorized cloud
1022 M. SANCHEZ ET AL.

computing into three service models: Infrastructure as of fourth main elements (people, data, processes, and
a service (IaaS), Platform as a service (PaaS), and data), in contrast with IoT, which consist of one part:
Software as a service (SaaS), which allow users to ‘things’. (Aazam and Huh 2016; Mohamudally 2017).
access software, operating systems, tools, and hard­ However, we consider that the definition of Cisco for
ware, as a service over the Internet. Shila et al. (2017) IoE is quite limited and must be extended to include
affirm that the IoS facilitates the organization of var­ the services dimension. For instance, in IoE, cloud
ious applications into interoperable services, as well computing services like storage services, clustering
as the use of semantics for the understanding, com­ services, among others, are commonly used in order
bination, and processing of data and information to outspread the limited capabilities of physical
from different service providers, sources, and formats. devices. Thus, by using Cisco’s IoE definition, there is
On the other hand, Vaquero et al. (2008) affirm that no place for services, because a service is not always
the main target of IoS is to present everything as a process neither data nor people or things.
a service on the Internet, including software applica­ On the other hand, a process (industrial or not) can
tions (SaaS), the platform to develop and deliver these be offered as a service using the XaaS concept of the
applications (PaaS (Exposito and Diop 2014)), and the cloud computing paradigm. In this sense, this paper
underlying infrastructure (IaaS (Vizcarrondo et al. proposes to enhance Cisco’s IoE framework by includ­
2012)). In that sense, everything as a service (XaaS ing the service dimension. In other words, IoE is rede­
(Perera et al. 2014)) refers to delivering anything as fined as the interconnection of (see Figure 3):
a service and includes the vast number of products,
tools and technologies that vendors can deliver to ● People: Humans behind a device using a human-
users as a cloud computing service. machine interface, a wearable device, or social
networks.
2.2.4. Internet of Things (IoT) and Internet of ● Data: Databases, unstructured data, or raw data
Everything (IoE) produced by things, services, or humans.
IoT is related to a network of physical devices and ● Things: It represents anything with connectivity
other items, with installed capabilities to allow the capabilities, like sensors, actuators, smartphones,
connectivity. (Chen et al. 2017a; Mezghani, Expósito, smart vehicles, computers, among others.
and Drira 2017a; 2017b). The authors of (Sengupta, ● Services: This is related to the XaaS model of
Gupta, and Vinayak 2017; Riggins and Keskin 2017) cloud computing, which means anything that
affirm that the purpose of IoT is to provide wireless can be accessed using a web service interface,
communication to various physical objects that we like a Database (DBaaS), Knowledge (KaaS),
use daily. On the other hand, Gupta et al. (2016) Software (SaaS), Business Processes (BPaaS),
expand this definition, saying that some of those among others.
mobile devices are semi-autonomous or autonomous
(smart), and can actuate in their surroundings to pro­ 2.2.5. System integration
vide services to users, who may or may not be in the Systems Integration is related to link together system
physical proximity of devices. On the other hand, components (Auger, Exposito, and Lochin 2017) like
Cisco defines the IoE as an era in which unprece­ software, hardware, or other systems and sub-
dented value is created by real-time interaction systems. Those components interoperate and provide
between actors, including not only things, but also solutions according to their goals (collective or indi­
people, processes, and data, all connected to the viduals) (Drăgan, Selea, and Teodor-Florin 2017). In
Internet (D. Lee, Choi, and Kim 2017). (Yang, Martino, the context of Industry 4.0, systems are usually inte­
and Zhang 2017; Martino et al. 2018; Shaikh et al. grated using technologies like IoT (Khan et al. 2017;
2017; Chen et al. 2017b) introduce IoE as the Mezghani, Expósito, and Drira 2017a; 2017b),
Internet that connects people, data, processes, and enabling the interoperability between things, data,
objects, giving connectivity of anything-anytime- people, services, and, and allowing them to connect,
anyplace for more intelligent health, smarter energy- communicate, coordinate, collaborate and cooperate.
efficient cities, smart transportation, among others. According to (Suri et al. 2017; Pisching et al. 2018),
IoE is considered as an evolution of IoT, which consists integration is given in three ways:
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1023

Figure 3. Actors of IoE.

● Horizontal Integration (inter-company integra­ integration needs of the manufacturing process.

tion): This is based on the collaboration or coop­ Coordination processes will allow a vertical integra­
eration between two or more companies, to tion between actors; however, cooperation and colla­
achieve individuals or common goals (Khan boration will allow integration in both sides
et al. 2017; Suri et al. 2017; Pisching et al. 2018). (horizontal and vertical).
● Vertical Integration (intra-company integration):
The vertical integration brings together the system 2.2.6. Modern human interactions
components within an enterprise, like production This sub-section refers to the ways people can interact
business process (BPaaS), applications (SaaS) with a system. In the context of Industry 4.0, Modern
devices, people, data, in order to allow them to Human Interactions try to answer the question: how can
coordinate, cooperate or collaborate (Khan et al. people interact with actors like devices, data, and
2017; Suri et al. 2017; Pisching et al. 2018). services?
● End-to-End Integration: The end-to-end integra­
tion mix the digital and real worlds in such a way 2.2.6.1. Augmented Reality (AR). AR is defined by
that real entities can interact with the cyber Kipper and Rampolla (2012), as the overlapping of
components of the system. For instance, devices virtual information on the real worldview. Besides,
that connect to the network can send informa­ Syberfeldt, Danielsson, and Gustavsson (2017) say
tion to the cloud, or people that communicate that applying this concept makes it possible to
with the system using a Human Machine enhance a human’s perception of reality. In the context
Interface (HMI) (Pisching et al. 2018). of Industry 4.0, operations less efficient and potentially
dangerous to technicians during their work might be
In this paper is proposed an incremental approach to driven using AR as an option to support the manufac­
deal with the integration challenges in Industry 4.0. turing processes (Pierdicca et al. 2017), and reduce risk.
This approach is called the 5 C integration stack and The usages of AR in Industry 4.0 are various and can be
consists of solving the issues on the specific layer of potentially applied to all activities taking place in the
the 5 C to which they correspond. In that sense, companies, like production, quality control, safety
connection issues must be solved in the first level. management, maintenance, and remote assistance,
Next, the communication issues must be dealt in training, logistics, and design (Pierdicca et al. 2017;
the second level. Once those actors can connect and Büttner et al. 2017). AR for Industry 4.0 provides many
communicate, the coordination, cooperation, and col­ advantages, such as design optimization, plant main­
laboration issues can be treated according to the tenance, and control, operator training, assistance and
1024 M. SANCHEZ ET AL.

resolution of incidents, among others, with the goals of its management (Khan et al. 2017). In this sense, Big
improving processes, reducing waiting times, increas­ Data deals with this problem in production processes
ing security and saving costs that are valuable charac­ by pre-processing the data generated mainly by sen­
teristics in Industry 4.0. sors, effectors, devices and people, looking for insights
and knowledge that allow the humans involved in the
2.2.6.2. Social networks. Cheung, Chiu, and Lee production process to make better decisions. In gen­
(2011) define social networks as ‘virtual communities eral, Big Data can be defined using the five ‘V’ as
that allow people to connect and interact over follows (Obitko and Václav 2015; Jirkovský, Obitko,
a particular subject synchronously or asynchronously, and Vladimír 2017; Chang and Wills 2016): Volume
or just to hang out together online.’ On the other refers to the large volumes of information that are
hand, (Liu et al. 2017) highlight social networks as generated daily. Velocity refers to the speed of how
sources of social and collective intelligence, from the data are produced and must be processed to meet
which is possible, for instance, to perform sentiment the demands. Variety refers to the different types of
analysis and to even create recommendation systems information formats, whether structured and unstruc­
(RS) based on people’s discussions and opinions. tured. Veracity refers to the reliability of the data.
Moreover, in Industry 4.0, the mining of social net­ Finally, Value refers to the meaning of the data in the
works could allow the manufacturing system to get operational context, that is, what is the real benefit that
useful information about people, in order to decide can be derived from them. Consequently, Big data
their preferences about a product, or their require­ analytics allows the collection and analysis of a large
ments, among others. number of data from different sources, in order to
support decision-making (Pierdicca et al. 2017). It is
2.2.6.3. Wearables. Basically, it is a computing fundamental in the context of Industry 4.0 to identify
device that people can wear (McCann and Bryson useful patterns, production models, (Santos et al. 2017;
2009). Besides, a wearable is an integral part of IoT Jirkovský, Obitko, and Vladimír 2017) as well as to ease
(Hao and Helo 2017), which allows tracking and mon­ the cleaning, formatting, transforming, and processing
itoring human activities (Mezghani, Expósito, and Drira the technical data (Khan et al. 2017). Mainly, techniques
2017b). Frequently, it uses Machine-to-Machine com­ like machine learning, text mining, data mining, pro­
munication (M2 M) to interoperate with other comput­ cess mining, service mining, semantic mining, and mas­
ing devices autonomously (e.g., smartwatches, glasses, sively parallel data processing like map-reduce and
or gloves). Mainly, Hao and Helo (2017) say that wear­ Hadoop, cloud computing, databases oriented to
ables can be helpful in the manufacturing industry for graphs and events, databases without schema, etc.,
increasing the human-machine interactions and for are necessaries for the analysis of the vast amount of
connecting them to the production process as human data.
resources. This section discussed the more familiar concepts
regarding the Industry 4.0 concept, which are essential
2.2.6.4. Semantic. According to (Obitko and Václav to adopt this new industrial revolution. The next sec­
2015; Grangel-González et al. 2016), semantic allows tion offers a state of the art around Industry 4.0, focus­
data or documents to be understandable by machines ing on the integration and interoperability challenges.
(computers, devices, services, etc.), by making their
meaning more explicit. In this sense, populating the
3. State of the art of industry 4.0 from the
data with semantic information will increase the inter­
integration perspective
operability between the actors of IoE, which is essential
to allow the coordination, cooperation, and collabora­ In this section, a selection of articles related to
tion processes, to reach their design goals. Industry 4.0 is presented. This state of the art is orga­
nized according to the actors/dimensions of the IoE
2.2.7. Big data analytics (people, data, things, services) versus the five levels of
In Industry 4.0, the integration of multiple manufactur­ integration, to identify the integration levels and
ing processes has generated a deluge of data from actors more studied. From our point of view, the 5 C
different sources, which requires new approaches for levels can be defined as:
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1025

● Connection: it allows entities of IoE to be linked

to the network and to share a standard media or
a channel for the communication. That means it
allows the actors to contact each other. The con­
nection is essential to allow communication.
● Communication: Once entities are connected, they
can exchange messages, allowing them to estab­
lish a conversation and interactions. Also, commu­
nication means that the entities can understand
each other. Connection and Communication are
essentials to achieve interoperability of the system,
as well as to allow the coordination, cooperation,
and collaboration processes. Figure 4. Existing works classified by the four actors of IoE vs.
the 5 C integration levels.
● Coordination: Basically, the coordination is an
activity carried out by a central entity (or orches­
trator) that allows coherently to harmonizing the corresponds to the number of papers dealing with the
execution of the tasks of a system (intra-systems integration of services. For instance, at the connection
integration or vertical integration). In terms of level, three research works consider the connection of
services, the coordination is closely related to things, the other three researches consider the connec­
the concept of intra-system orchestration (inter­ tion of data, three the connection of people, and two
nal to a business process or system). researches the interconnection of services.
● Cooperation: It consists of a negotiation process Furthermore, at the coordination level, two
that allows the entities of the same system (intra- researches worked with the coordination of Things,
system integration or vertical integration) or entities only one paper considered the coordination of people,
of two or more systems (inter-systems integration and one research is about the coordination of services.
or horizontal integration) achieving agreements in Also, in the literature review, it was found only one
the execution of their tasks, in order to accomplish paper dealing with the integration challenges in more
individual objectives. Cooperation is related to than one level. It means that any solution provided by
inter-system orchestration; however, it does not previous researches can allow all actors to connect,
rely on a central coordinator. communicate, coordinate, cooperate, and collaborate.
● Collaboration: refers to entities of two or more In the next sub-sections, these research works are
systems (inter-system) that work together in analyzed according to the 5 C level that best suited to
order to achieve a common goal that partici­ each work. The works have been selected because
pants would not be able to accomplish working they cover several aspects of the proposed classifica­
alone. Collaboration is related to inter-system tion model, or they have some relationship with the
choreography (interactions between autono­ problem of autonomous integration in Industry 4.0,
mous processes). Collaboration does not rely on mainly for the highest 5 C levels (coordination, coop­
a central coordinator. eration, and collaboration). In this classification, it can
be noticed which integration level has been in recent
According to this classification, a survey of research years the most studied, as well as the actors involved
works mainly focused on the levels of integration of at each integration level.
the IoE actors was elaborated. In that sense, Figure 4
shows a summary of the recent researches organized
3.1. Connection
by the integration levels. In this case, a blue bar indi­
cates the number of research works dealing with the Research works positioned at this level focus mainly
integration of Things, an orange bar means the number on the interconnection of entities, using a shared
of research works that focus on the integration of Data, medium, but without defining explicit mechanisms
the gray bar specifies the number of researches that for communication, coordination, cooperation, or
considered the integration of people, and a yellow bar collaboration.
1026 M. SANCHEZ ET AL.

Molano et al. (2017) describe an architecture for IoT layer converts data to a unified semantic form, accord­
applied to the industry (denoted as IIoT, for Industrial ing to SHS ontology (it corrects the damaged data if
Internet of Thing), which integrates IoT, sensors, needed).
actuators, social networks, and cloud computing. Moreover, on this layer, the semantic heterogeneity
The prototype architecture contains five layers. The is solved, and RDF triples are created in order to popu­
Sensing layer comprises several types of devices, and late the data with semantic information. The Data
it is responsible for collecting data from sensors or storage layer is in charge of storing the triples in the
other devices, as well as to manage the manufactur­ RDF format. Finally, the Analytic layer provides direct
ing and logistics processes. The Database layer con­ access to the storage layer for analysis tasks or custo­
tains the physical (SQL and NoSQL databases) and mizing the user queries. The authors affirm that the
virtual databases (logical links to the databases in main advantages of the integration using the SHS
the network nodes). The Network layer support all ontology is that the ontology describes the reality in
the infrastructure (physical devices), allowing devices its representation, and data can be easily queried in
to connect using wireless or wired networks. SPARQL. Same as previous works, this research is
Moreover, the network layer allows sharing the focused on put data available to other actors and deal­
data with other devices connected to the system, ing with the data heterogeneity issues. They allow the
enabling the interaction between the Sensor layer connection between actors. For that reason, this paper
and the User layer. Similarly, the Data Response layer classifies it as belonging to the connection level.
represents a data set whose goal is to keep the per­ Notably, (Molano et al. 2017; Jirkovský, Obitko, and
sistence of other layers. The User layer contains the Vladimír 2017) are relevant for our research because
API used by ERP applications, in order to monitor the they describe how to connect actors of Industry 4.0
raw material, the equipment failures, the quality con­ through data. Other works in this domain allow the
trol and programming of the production. This layer is inter-connection of actors using techniques like Big
a Middleware that provides several services, such as Data (Khan et al. 2017), AR (Pierdicca et al. 2017;
data compilation, transmission, data processing, IoT Syberfeldt, Danielsson, and Gustavsson 2017), Network
services, etc. Molano et al. (2017) focused on the protocols like TCP/IP (Bohuslava, Martin, and Igor 2017;
interconnection of data with things through services, Exposito 2013), or works that present architectures for
which access databases in the cloud computing and CPS integration (J. Lee, Bagheri, and Kao 2015).
make data available to each device. It means that
devices are connected through the data. However,
3.2. Communication
this work does not include the coordination topic
between the integrated actors. Because of that, this Most of the studied works are classified at this level
paper considers it to be positioned between the con­ because they allow entities to communicate, making
nection and communication levels. abstraction of underling connection details, and with­
On the other hand, Jirkovský, Obitko, and Vladimír out proposing explicit processes for coordination,
(2017) use Semantic Web technologies mixed with cooperation, or collaboration.
a Big Data approach for data integration for CPS. Román-Ibáñez, Jimeno-Morenilla, and Pujol-López
They use a Semantic Web approach to deal with the (2018) propose a communication layer aimed to retrieve
semantic and platform heterogeneity issues. The data of robotic arms manufactured by different firms.
authors propose an ontology (called SHS ontology) The proposed system allows monitoring the status of
for the description of the industrial data. The SHS robotics cell in a footwear factory and displays a 3D visor
ontology contains structures that allow modeling dif­ that shows a simulation of the movements of robotic
ferent observations: the physical qualities, the units of arms. Essentially, this research work unifies the different
measurements, or the external data sources. The archi­ communication protocols used by manufacturers into
tecture is divided into four parts. First, the Data acquisi­ a single one. Moreover, they have defined a custom
tion layer collects data from sensors, other systems communication protocol over TCP/IP to retrieve the
(MES/ERP systems), and some relevant external data data from the monitoring system. This protocol supports
sources. Besides, this layer is in charge of solving plat­ 128 types of messages, but only five have been defined
form heterogeneity issues. Next, the Transformation (MSG_TIME, MSG_NOTIFY, MSG_DOF, MSG_JOINTS, and
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1027

MSG_SELECT). MSG_TIME is used to maintain the time­ Reporting, etc. The seventh layer applies mechanisms
line of the monitored data. MSG_NOTIFY allows sending for Security, Administration, and Monitoring.
notification messages from servers to clients. MSG_DOF Moreover, this layer includes components needed
is used to change the number of degrees of freedom of in the other layers in order to guarantee the proper
the robot arms’ joints. MSG_JOINTS allows maintaining operation of the whole infrastructure. In general, this
the angle value of each joint in the robotic arm chain. work allows the data to be available to other actors,
Finally, MSG_SELECT allows a client to collect data from letting them communicate indirectly through it.
a specific robotic arm. The communication messages are However, this research does not study coordination
protected from unwanted attackers. Thus, all the data processes, nor proposes processes for cooperation
shared between servers and clients is encrypted to avoid and collaboration directly. For this reason, it is con­
man-in-the-middle attacks. This paper corresponds to sidered to be positioned at the communication level.
the communication layer due that its goal is to allow Also, it considers only the data dimension of IoE.
robotic arms from different manufacturers to commu­ Moreover, even if the author claims this works are
nicate using a standard protocol. adapted to Industry 4.0 (Santos et al. 2017), they did
Santos et al. (2017) propose a Big Data Analytics not present a precise application on this domain.
architecture that includes layers dedicated to deal Similarly, Suri et al. (2017) provide a solution for the
with data needs, from the collection to the analysis modularity and interoperability issues related to
and distribution. The proposed architecture is divided Industry 4.0 from a systems integration viewpoint,
into seven layers. Components define each layer, and focusing on the ‘vertical integration’ of system using
each component can be associated with some tech­ the model-driven engineering (MDE) approach. This
nological tool. The first layer represents the Big Data approach enables heterogeneous systems to commu­
producers and consumers’ entities; these entities are nicate in a low-coupled manner. In particular, this
usually consumers of raw data, indicators, or metrics. approach is oriented to industrial robots, which per­
The second layer (Data sources layer) represents the form standard repetitive tasks. The communication
different sources of data, including components such model consists of two layers. The model-based beha­
as Databases (operational/transactional databases), vior layer, in which the task execution model is cre­
files, ERPs, E-Mail, sensors, among others. All this ated using an activity diagram in UML 2.0; and the
data will feed the ETL layer (extraction, transforma­ robot’s implementation layer, which is in charge of
tion, and loading processes). The third layer corre­ transforming the activity diagram designed in the
sponds to the process of extracting data from data previous layer into instructions recognized for robots
sources and storing it into the Big Data Warehouse by using some available frameworks designed for this
(BDW), using several technologies to implement the purpose (i.e., Papyrus). The execution of the robots is
ETL process and to integrate data from multiple data made using API calls through the execution model
sources. The fourth layer is the Data Storage layer. This (on-line execution), rather than deploying the source
layer was divided into two sub-layers, which contain code on the system (off-line execution). In this sense,
different components that must be used according to this approach allows the creation of complex systems
the context. Consequently, the data storage sub-layer of sensors and actuators, with low computational
stores data into a NoSQL database like Cassandra or power and low energy usage.
HBase-streams in real-time. Consequently, the main objective of the previous
On the other hand, The Hadoop BDW sub-layer is in research is to allow things to communicate transpar­
charge of preserving the historical data. Once the data ently and in a loose-coupled way. Due to this reason,
was stored in the BDW, it will be available for data it has been classified in the communication level. This
analytics through the SQL query engine. The fifth paper is relevant for our research because it shows
layer, the Raw Data Publisher, enables access to the how to integrate and communicate actors of Industry
data by providing Web Services for the data stored in 4.0 using an MDE based approach, to send orders to
the Data Storage layer. The sixth layer, the Big Data devices.
Analytics, includes components to facilitate the ana­ On the other hand, Bohuslava, Martin, and Igor
lysis of the data, making available different data ana­ (2017) enable communication based on the standard
lysis techniques like Data Visualization, Data Mining, Ethernet (IEEE 802.3) for the control of the robotic cell.
1028 M. SANCHEZ ET AL.

The communication protocol used in this model of cyber level uses a machine–cyber interface (CPI) in
production robot cells is TCP/IP. The Control applica­ order to allow the interconnections between
tion is divided into several parts, consisting of a server machines. On the other hand, the cognition layer
running the control task, and some client subpro­ generates detailed knowledge of the monitored sys­
grams running on each robot. All the requirements tem, and makes it available to experts, allowing them
of the cells, as well as the instructions from the control to make the correct decisions. According to the
center, will be processed as TCP/IP sockets. Complete authors, this level requires proper user interfaces/
communication with the sockets takes place only dashboards in order to transfer the acquired knowl­
through the central unit, which coordinates the com­ edge to the users completely. The configuration level
munication among the robot cells. This coordination allows self-configuration and self-adaptation of the
is based on the messages exchanged thought the devices, by getting feedback from cyberspace to phy­
sockets according to two variants: a) Confirmed coor­ sical space, acting as supervisory control. That config­
dination, where the completion of each operation is uration allows applying the corrective and preventive
notified to the central unit b) Unconfirmed coordina­ decisions (defined in the cognitive level) to the mon­
tion, in this case, the continuity of the activity is not itored system. In that sense, this system acts as
conditioned to receive a confirmation message from a resilience control system (RCS). This research pre­
a superior object. In this sense, robot cells can com­ sents an intelligent middleware, which allows the
municate using the TCP/IP protocol, with a central coexistence of devices and people, facilitating the
unit making the coordination of the whole commu­ collection and transformation of data between them.
nication process. This paper deals with problems of Due to that, this research is positioned at the level of
connection, communication, and coordination of communication. Consequently, this research work is
robotic cells; however, the authors of this research relevant for our research because it describes how to
work said that the primary goal of it is to allow deploy a CPS and how to communicate and share
those robots to communicate. It is the reason why information among things.
this paper considers it to be in the communication Sanchez et al. (2018b) have developed a solution for
level. This paper is essential for our research because the interoperability between MAS and the Software
it describes how to incorporate an existing low-level Oriented Architecture (SOA) paradigm for intelligent
communication protocol in new devices, to allow environments (MAS-SOA Integration). Mainly, Sánchez,
them to communicate appropriately. Aguilar, and Exposito (2018b) allow bidirectional com­
Additionally, Lee, Bagheri, and Kao (2015) propose munication between the agents of a MAS and the web
a unified 5-level architecture as a guideline for the service deployed in cloud computing, to give support
implementation of CPS. The proposed 5-level architec­ to the actors of an intelligent environment. Principally,
ture provides a step-by-step guide for developing and the agents of the middleware characterize people,
deploying a CPS in manufacturing environments. The devices, and services involved in the intelligent envir­
first level, the Smart connection level is in charge of onment, which interact and exchange data. Cloud
acquiring the data directly from the sensors, or of computing services process this data in order to let
collecting it from the controller or the enterprise man­ agents using it and make decisions oriented to sup­
ufacturing systems, such as ERP, MES (Manufacturing port the activities of the users in an intelligent envir­
Execution System), SCM (Software Configuration onment. The critical aspect of the solution provided by
Management) and CMM (Capability Maturity Model). Sánchez, Aguilar, and Exposito (2018b) is the incor­
The second level, the Data-to-information conversion poration of an SOA-MAS communication sub-system,
level, infers useful information from the data using which is in charge of transforming messages from web
several tools and methodologies. The Cyber level acts services to a language that agents can understand,
as a central information hub. The information is sent such as FIPA-ACL, and vice versa. Also, in Sanchez,
from every connected machine in the network to the Aguilar, and Exposito (2018a), the fog computing para­
Cyber level. digm was added to the solution proposed in Sánchez,
Moreover, specific analytics tasks are used to Aguilar, and Exposito (2018b), to avoid the issues of
extract additional information that provides better the cloud computing-based solutions. In that sense, in
insights regarding the status of each machine. The Sanchez, Aguilar, and Exposito (2018a), the authors
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1029

combine the fog-computing paradigm with the MAS- step, the data source associated with each indi­
SOA integration sub-system in order to solve issues of cator is defined.
geolocation, real-time, and low-latency. Mainly, these (4) Link equipment of new and old machines.
works are essential in the Industry 4.0 context, because Determine the proper equipment required for
they allow the interoperability of actors in any cloud each machine being automated in order to
computing-based intelligent environment, like smart allow it to operate autonomously.
cities, smart factories, etc. Moreover, Sánchez, Aguilar, (5) Create standalone networks. The machinery
and Exposito (2018b) and Sanchez, Aguilar, and might communicate using a dedicated and
Exposito (2018a) are focused on solving issues related independent network in order to avoid com­
to the communication layer, which allows the actors to munication conflicts.
exchange information among them. (6) Generate alerts when processes’ variation is
Other works in this domain allow explicitly or impli­ detected. An alarm must be generated when
citly communications between actors, applying the system encounters a fault, such as the shut­
a variety of techniques, like MAS (Romero et al. down of a motor, actuators, among others.
2017), IIoT (Molano et al. 2017; Wan et al. 2016), Big (7) Improve feedback and follow up processes. The
Data (Jirkovský, Obitko, and Vladimír 2017), human- production orders in which operators are work­
robot interaction (Huber and Weiss 2017; Nelles et al. ing, as well as operation’s errors and problems,
2016), AR (Pierdicca et al. 2017; Longo, Nicoletti, and are captured using a data collection software
Padovano 2017), or Middleware (Ferrera et al. 2017). and send to an ERP software, such that all the
information about the process can be available
when it is requested for failure diagnosis.
3.3. Coordination
(8) Test and validate the system. Verify if the sys­
At this level, this work considers only those paper that tem operates correctly or needs to be tuned up.
allows coordination process using a central coordinator,
according with the requirements defined previously (see The case study for Orellana and Torres (2019) was
section II) for the coordination level (centralized intra- conducted in an enterprise where the machinery
system coordination). In that sense, Orellana and Torres had more than 47 years of operation. The results
(2019) propose a procedure to transform a legacy man­ show that after the production process was trans­
ufacture process into a smart factory level 2, according formed into a smart factory level 2, the production
to Industry 4.0. Essentially, this proposal allows vertical line was able to reduce the average production time
integration, which guarantees the actors involved in the from four days to three hours. This paper is positioned
internal production process to share information. at the coordination level because it improves the
Notably, the remarkable point of this proposal is that it coordination of the whole production process.
grants integration without buying new machinery. Soto, Tavakolizadeh, and Gyulai (2019) present an
Moreover, their proposal uses Industrial IoT to achieve orchestration framework that combines IoT and
its integration goals. The procedure comprises eight machine learning for failure detection in production
steps that are executed continuously until the industrial lines. The solution comprises three fundamental ele­
process works correctly. ments. Firstly, the production line is exposed using an
IoT Connector that is responsible for transforming the
(1) Define management indicators. Define the indi­ data from different production protocols into network
cators that will be used for evaluating the pro­ protocols or message queue systems. Secondly, the
cess. The authors propose to use the ISO 22,400 connector propagates the data using a Broker accord­
standard (Kang et al. 2016) for this purpose. ing to the IoT standards. Thirdly, the data is processed
(2) Define the process’s inputs, signal, and sensors. using a learning agent that orchestrates all the compo­
This step helps to determine which measure­ nents’ behavior and selects the learning algorithm
ment instruments the machinery requires, and depending on the data characteristics and the current
how to link them with other machinery’s sensors. use case.
(3) Identify and choose data sources based on the Moreover, this component uses python to build
corrective and preventive maintenance. In this the machine-learning model for decision-making.
1030 M. SANCHEZ ET AL.

This framework was evaluated holistically, using linked together in a specific order. At the end of the
a realistic simulation. However, in a real production assembly phase, the application makes a verification
line, issues not covered by this solution might hap­ of some parts of the final object. In this sense, the
pen. This research work is positioned at the coordina­ application displays the assembly instructions one by
tion level because it can orchestrate the behavior of one using the head-mounted display (HMD) that the
the production line. However, there is still a significant user wears. Moreover, the application uses textual
amount of work to do in order to promote autono­ information and 3D models of the real scene in
mous interoperability. order to help the operator to finish the task quickly.
Ivanov, Sokolov, and Ivanova (2016) introduce the That means that for implementing this application,
dynamic control concept and a dynamic model to some 3D models for each real component must be
coordinate activities in cyber-physical supply chains, created; those models must keep the same dimen­
based on the smart manufacturing concept. The sions and components as the real scene.
authors propose a scheduling approach based on Moreover, the 3D model of the real object is cre­
making a temporal decomposition of the scheduling ated using different colors for each component to
problem to allow the dynamic execution of the jobs. allow users to recognize them easily. This research
They use a dynamic structure control (SDC) approach connects people with things in a coordinated way,
model, which is a dynamic interpretation of planning, in which the android app is the coordinator that dis­
in concordance with the execution time. Additionally, plays information through the HMD interface to allow
SDC is combined with the optimal program control users to assemble an object more comfortably. This
(OPC) theory and mathematical programming (MP). paper is positioned at the coordination level because
The dynamic of planning is because the decisions on it allows achieving a global goal using a central coor­
supply chain planning are taken for specific intervals dinator. Moreover, (Pierdicca et al. 2017) present an
of structural constancy. In this sense, a static optimi­ unusual approach that combines AR with devices and
zation problem is solved with the help of MP for each people to enable coordination, which can be very
time interval, while OPC is used in order to define and useful in the context of Industry 4.0.
to model the transitions between the time intervals. Another research work with some level of coordina­
Moreover, the supply chain is modeled mathema­ tion is (Bohuslava, Martin, and Igor 2017); however, that
tically as a networked system described through con­ work is focused mainly on the connection of things.
trol models M1-M2 (schedule for material supply
processes, schedule for services, respectively). Then,
3.4. Cooperation
when the manufacturing process starts, the M1-M2
process assigns services to business operations in The papers positioned at this level promote a specific
sequential order. Next, M2-M3 (M3: schedule for mechanism for cooperation. For example, Huang et al.
resources) assigns and schedules services to informa­ (Huang et al. 2017) propose a community energy
tion resources. Finally, M3-M4 (M4: schedule for infor­ system planning (CESP) model based on a Multi-
mation systems modernization) is launched in order Agent Systems (MAS), in order to improve the energy
to reconfigure the system. The coordination happens utilization within a specific community. In the solu­
in the system’s interconnections; for instance, the tion, each participant is viewed as an agent.
output of M1 is used in the constraints of M2; Furthermore, the stakeholders are represented as
Analogously for M2, M3, and M4. This research is CESP agents too. Four types of stakeholders were
positioned at the coordination level because considered into the model: Governments, People,
a central entity executes the coordination process. Energy firms, and Energy facilitators. Additionally,
Also, this paper is essential for our research because the spatial location is also considered into the
it shows how the coordination processes can be model, due to the transmission cost of hot and cold
deployed autonomously in the industrial domain. water. In this sense, all agents are organized in
Pierdicca et al. (2017) develop an AR Android appli­ a spatial hierarchy and divided into different groups
cation, in the context of Industry 4.0. This application based on whether they have similar interests or not.
assists an operator in order to allow assembling an The negotiation process is only performed between
object composed of several components that must be agents of the same group, in order to improve the
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1031

negotiation efficiency and to reduce the negotiation machine) that wins the bid will execute the task. After
time. The information needed for the negotiation pro­ the smart machine agent is selected, it is necessary to
cess, like price, supply and demand, policies, and other start moving the smart product from the current posi­
agents’ initial planning, etc., is available to all agents. tion to the target one, by constructing a conveyor belt
That is, they share their belief-desire-intention data route chain. Again, a new negotiation round is required
within the group. Mainly, this research allows agents in order to create the route chain from the start point to
to build a decision-making system, which lets agents the target point. This research is focused on the coop­
performing negotiations in order to improve the erative process to organize smart products and smart
energy consumption within a community. Thus, each devices into a smart factory, by using the MAS para­
agent of the platform can find potential partners, digm where each agent has individual objectives. This
negotiate and construct its decision-making model, research is classified at the level of cooperation because
with the primary goal of making an optimal decision the process of coordination is not centralized, like was
related to the energy consumption within the com­ defined in section 2. Besides, this work is relevant for
munity. The information needed by the agents to our research because it gives us a vision of how smart
build their decision-making model includes databases, objects can drive a cooperative process.
as well as negotiation models (such as persuade, threa­ Other works in this domain propose cooperation
ten, inducements, and promise). Every agent uses processes using technologies like MAS (Romero et al.
a specific format that the other agents know to upload 2017), model-driven engineering (Suri et al. 2017), or
the negotiation models. However, in order to guaran­ human-robot interactions (Nelles et al. 2016).
tee the privacy of the data, private information is only
visible to interrelated agents. These agents cooperate
3.5. Collaboration
without a central coordinator, in order to achieve the
objectives discussed previously; because of that, it is In this sub-section, work focused on the study of the
classified at the cooperation level. In the same way, explicit collaborative process is presented. That
this paper shows how enterprises can offer their ser­ means interactions between system actors in order
vices as cloud computing services, and how customers to achieve a common goal. In that sense, Romero
can select the service with the best benefits autono­ et al. (2017) propose a social factory architecture
mously. Particularly, this paper is significant for our based on adaptive, collaborative, and intelligent
research because it shows how the MAS paradigm MAS. Moreover, they explore the role of what they
can be used to make autonomous decisions and name a social operator 4.0 in the context of smart and
achieve a specific grade of efficiency cooperatively. social factory. Mainly, people, devices, and software
Also, D. Li et al. (2017) focus on developing a MAS systems socialize together (cooperate or collaborate)
that can deal with the complexity of the cooperative in real-time, to support manufacturing and services
processes in smart manufacturing, using a structure operations. The authors define a social operator 4.0 as
consisting of intelligent agents with cloud computing- a type of operator that combines smart wearable
based feedback and coordination assistance. The nego­ solutions in conjunction with advanced human-
tiation mechanism allows a smart product (instantiated machine interaction (HMI) technologies to promote
as an agent) to act as manager, while the smart cooperative/collaborative processes with other social
machines and smart conveyor belt (instantiated as operators, social machines, and social software sys­
agents too) acts as contractors, competing by tasks. In tems. The MAS is used in order to simplify the com­
this case, an RFID tag is used for the communication of munication between the cyber-physical elements,
the agents, by reading/writing it. Smart machine agents that means, humans, machines (real world), and soft­
or smart conveyor belt agents initiate the negotiation ware entities, as well as to distribute tasks (based on
process. The smart product agent publishes a task, and their competencies) and share & trade control in col­
the smart machine agents receive it, deciding whether laborative tasks.
to bid for the task or not. The product agent uses a rule- Moreover, the MAS keeps as much as possible
based decision system in order to calculate the perfor­ human inclusiveness within the manufacturing pro­
mance indicators that allow deciding which contractors cess, without compromising production goals and
are awarded or rejected. The contractor (smart efficiency. Finally, the MAS can improve the skills of
1032 M. SANCHEZ ET AL.

the human and machines through learning and prac­ Consequently, the goal of the final phase is to
tice, for what it must record and track their evolution. define a concrete plan addressing the require­
The social factory MAS is composed of human agents ments and problems detected for each challenge.
that characterize humans and their skills; artificial Finally, this paper is a proposal for a strategic
agents (machines) that characterize the machines method for collaboration among industries. The
and their capabilities; interface agents that character­ authors of (Riel and Flatscher 2017) have designed
ize interaction rules and conditions for assisting a use case to demonstrate how this proposal fits in
humans and machines interfacing with the rest of Industry 4.0. However, this is a manual process that
the system. Similarly, Broker agents characterize the does not involve any kind of automation. This
levels of automation available in the system and the research is fundamental for our purposes because
rules for sharing and trading control in human- it shows a method that could be automatized to
machine cooperation, in order to efficiently allocate allow autonomous collaboration processes in
and distribute tasks between the cyber part and the Industry 4.0.
humans at the workstations of the manufacturing Likewise, Richert et al. (2016) make an empirical
system. This research provides a suitable mechanism study of the collaborative problem in human-robot
for facilitating collaborative tasks in smart factories, -teams. The method uses virtual reality in order to
due to this reason, it is positioned at the collaboration simulate a task that can only be accomplished
level. This work is relevant for our research because it through teamwork between robots and humans.
shows how a MAS can be an excellent solution to deal The participants are immersed in a virtual scenario
with collaboration issues autonomously. developed in Oculus Rift (a virtual reality system).
On the other hand, Riel and Flatscher (2017) deal The chief objective of this experiment is to analyze
with the creation of a structured methodological how the appearance of robots affects teamwork
approach to strategic production planning (SPP), outcomes. In that sense, a humanoid robot is
intending to establish an integrated manufacturing used in one experiment and an industrial robot in
road-mapping process. The essential idea is that the another. The human forms a team with each robot,
stakeholders involved in an integrated design process and his physical reactions are studied while the
shall run a series of stages of divergent and conver­ team develops a task collaboratively. A set of
gent thinking. At each stage of divergence, the ideas instructional commands is provided beforehand
about the design process are generated out-of-the- to the human, in order to allow the communica­
box thinking. Then, in the convergence stage, the tion with the robot.
ideas are consolidated and evaluated, with the pri­ On the other hand, Oculus Rift collects all the data
mary goal of deciding how to proceed with every generated during the study. The authors suggest
single idea. Typically, multiple parallel paths are cre­ using big data analytics in order to understand the
ated, where each path represents a particular set of nature of the collaborative process better; however,
ideas being worked. For the particular design pro­ they consider that these experiments are only the
blem of the SPP, the authors proposed a schema stepping-stone to much more detailed research.
with three phases (each phase involves divergence Although the authors of (Richert et al. 2016) say that
and convergence stages). The first phase is in charge this research is linked to Industry 4.0, they do not
of identifying the most relevant topics, based on the clarify how this method can be used in this context
use of techniques like brainwriting, extreme scenar­ to increase production, reduce costs, among other
ios, etc. In the second phase, the participation of the things. Essentially, our interest in this paper is because
top management representatives is required to prior­ they measure the impact of mixing robots and people
itize the topics, according to the company strategy. collaboratively.
The third phase uses the ranked topic list as input, to The next section presents a discussion regarding
deal with the specific challenges linked to each topic. the integration challenges in Industry 4.0, as well as
Next, the experts are involved in the process to iden­ a summary of the aspects covered in the solutions
tify and group each selected topic focusing on what presented in the current section at each integration
has to be done rather than on when. level.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1033

4. Discussion ● The definition of different integration styles,

horizontal (between companies), vertical (intra-
4.1. Integration challenges in industry 4.0
company), end-to-end (mixing digital and real-
Industry 4.0 is a concept still in development, so there world), etc. (Huber and Weiss 2017; Terán,
are many challenges to solve. Figure 5 shows the Aguilar, and Cerrada 2017; Yang, Martino, and
main current challenges of Industry 4.0 and their Zhang 2017). Mainly, horizontal integration is
relationships with the 5 C levels. It can be seen that needed in order to allow collaboration or coop­
the Standardization challenge is presented at all eration among enterprises. Vertical integration
levels. Also, in the last three levels, there are more is also required to allow coordination, coopera­
challenges related to planning, decision making, and tion, and collaboration between actors involved
negotiation, that are oriented to allow actors to in the production processes, and the End-to-end
autonomously interoperate in order to coordinate, integration is required to allow people to parti­
cooperate, and collaborate during the manufacturing cipate in the production process in a more nat­
processes. Those challenges are detailed as follows: ural and less invasive way.
● The specification of negotiation and conver­
● The standardization throughout the 5 C integra­ gence mechanisms must be proposed with the
tion levels means that standardized protocols, intention of allowing actors to achieve agree­
data & format, etc. are essential to allow a good ments in the execution of their tasks.
understanding of the information shared by the Particularly, some works are proposing auto­
actors, to allow autonomous 5 C processes. nomic negotiation mechanisms for the 5 C

Figure 5. Integration challenges in industry 4.0 by integration level.

1034 M. SANCHEZ ET AL.

highest levels of integration (coordination, coop­

eration, and collaboration), to autonomously
integrate the diverse actors that take place in
Industry 4.0, and especially, data, people and
services (see Figures 4 and 5).
● The definition of security and privacy of data
mechanisms to allow a robust integration of the
actors (Molano et al. 2017; Jirkovský, Obitko, and
Vladimír 2017; Huber and Weiss 2017). It is essen­
tial to provide some mechanism that preserves
the security and the privacy of the data in order
to avoid an unwanted attacker to gain access to
the information of the production process.
● The management of the heterogeneity of data Figure 6. The research works by integration level.
and actors (X. Li et al. 2017a; Molano et al. 2017).
Different actors generate a large amount of data, interconnect all the actors, leaving many challenges
information, and knowledge in different formats still to be addressed, for instance, security, semantic,
that should be semantically integrated in order to integration, and heterogeneity. From the case study
allow actors to understand the messages correctly. can be noticed that connecting actors are
● The management of the complexity of the plan­ a significant feature, in order to allow them to commu­
ning, to allow self-configuration of the production nicate; that means, allow people, services and things to
process (X. Li et al. 2017a; Huber and Weiss 2017). have access to the data, allow people to have access to
● The management of the decision-making process services or to connect to the system using a device, etc.
(Molano et al. 2017) to make optimal decisions On the other hand, at the communication level,
related to fit the goals of the manufacturing research works are focused on allowing interopera­
process. tion of actors by using a variety of methods, some of
them use a service-oriented architecture, while others
allow communication using low-level protocols. In
4.2. Studied researches
the same sense, Figure 4 shows that some works
The research works presented in Section III shows focus on communicating things and people, while
some of the main proposals addressing the integra­ others are focused on communicating data and ser­
tion challenges in Industry 4.0. However, there is not vices. Furthermore, it does not exist a complete solu­
a complete solution that covers all the integration tion that allows all the actors to communicate
aspects through the 5 C levels in a production pro­ correctly. The communication between actors is fun­
cess (see Figure 4). Figure 3, illustrates that all actors damental to make possible the 5 C highest levels.
in the production process must be able to connect, From the case study, if the actors are not able to
communicate, coordinate, cooperate, and collabo­ communicate, then the production process will be
rate in order to achieve the production goals effi­ limited because coordination, collaboration or coop­
ciently. However, Figure 6 shows that the most eration processes would not be possible in an effi­
studied integration level is communication, followed cient way, and the actors will not know what, how and
by the connection and cooperation levels, leaving at when to perform a task.
the last place the collaboration and coordination At the coordination level, most of the research
levels. works deal with this topic mainly following an implicit
At the level of connection, the presented studies try approach, and only services and things are explicitly
to connect the actors, using technologies like Semantic coordinated, leaving much work to perform in this
Web, Big Data, among others, as connection medium. area. However, as illustrated in our case study, smart
However, as can be seen in Figure 6, more research production asks for automatic discovery and coordi­
works are focused mainly on the interconnection of nation with an emphasis on allowing autonomous
things, and there is not a solution that allows to coordination of production processes.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1035

On the other hand, the level of cooperation has 4.3. Integration technologies
been moderately worked (not as much as the connec­
Figure 7 shows a summary of the commonly used
tion and communication levels, see Figure 6) using
technologies that allow the integration and interoper­
techniques such as human-machine interfaces or
ability of actors in Industry 4.0.
MAS. However, a complete solution covering the
At the connection level, it can be seen that the
cooperation of all the actors does not exist, leaving
technologies used to allow actors to get in touch
still much work to do at this level. One example of
between them are: Augmented Reality, Industrial
these works is how to make decisions or plan new
Internet of Thing, Semantic Web, and TCP/IP. It means
activities while everything is cooperating. In that
that connection not necessarily indicates a network con­
sense, our case study illustrated the importance of
nection; it implies that actors can contact each other.
offering cooperation between internal or external
E.g., Augmented Reality makes it possible to connect
actors, allowing turning a traditional production pro­
people with other actors of a production process; how­
cess into a smart production process. In such a way,
ever, it can also allow communication or coordination.
the smart production process can reason in order to
Besides, as it was discussed in the previous sub-
discover the raw materials needed to complete the
section, communication was the level most studied
orders.
by researchers in past years. It explains the variety of
Similarly, at the collaboration level, several works
technologies used at that level, as shown in Figure 7.
have been presented, but less than in the coopera­
Starting from low-level protocols like TCP/IP, medium
tion, connection, and communication levels (see
level protocols like FIPA, until more high-level proto­
Figure 6). The proposed solutions mainly include
cols such as the service-oriented one. Furthermore,
mixing of HMI technologies with MAS, SPP,
technologies as Big Data Warehouse allows communi­
among others; studies on specific scenarios and
cation indirectly, by letting actors getting and sending
possible solutions are proposed, in order to allow
information through data, that also can be used for
collaboration, mainly of things, and in some cases
future coordination and learning analytical processes.
of people. As can be seen from our case study,
Accordingly, at the Coordination, Cooperation, and
collaboration is fundamental in the production pro­
Collaboration levels, it can be noticed that the techni­
cess, in order to allow actors to achieve the general
ques are more related to AI. This work believes it is due
goals of the manufacturing process. However,
that on those levels, actors need to incorporate nego­
enabling autonomous cooperation (by discovering
tiation and convergence protocol in order to resolve
or creating a collaboration plan) is essential in the
conflicts in the execution of their task. Moreover, actors
context of Industry 4.0, to increase the efficiency of
must deal with the complexity of planning and deci­
the production process.
sion-making, and AI helps to solve those challenges.
From Figures 4 and 5, it can be deduced that there
is a need for an integration solution involving all
actors along all levels of 5 C. In particular, this work
4.4. Case study
thinks that most of the integration issues should focus
on the heterogeneity of actors that take place in To illustrate the results of this research, we will use the
Industry 4.0, which need to be able to communicate case study described by Vachálek et al. (2017). The case
in different ways and to process large and heteroge­ study corresponds to a production line manufacturing
neous amount of data, information, and knowledge. of pneumatic cylinders. This production line consists of
In the same way, the autonomous coordination, six stations: Distribution, Test, Processing, Handling,
cooperation, and collaboration processes need to be Sorting, and Assembling. At the first station, the opera­
further studied in the context of the Industry 4.0 con­ tor supplies the line (put the components to the system),
cept. In particular, semantic integration between actors according to the production plan. An arm takes the
to allow sharing a common understanding is required. components from the tray and put it to the next station.
In Summary, there is not a complete solution that The second station is the test, which checks the size of
allows the integration of all actors of Industry 4.0 and each component. The process station performs the dril­
enables autonomous mechanisms for coordination, ling and control the size of the hole after drilling. At the
cooperation, and collaboration. handling station, a manipulator puts each processed
1036 M. SANCHEZ ET AL.

Figure 7. Commonly used integration technologies by level.

component into the sorting section, on which these (3) At this point, we found the integration chal­
components are sorted in different conveyor belts, lenge, because it is necessary to put all the
according to the type of piece being produced. Finally, actors in the production line to share infor­
in the Assembling station, an operator combines the mation and act autonomously, in order to
components (the piston, the spring, the cylinder body, produce each kind of cylinders. This process
and the lid), and packed them with the corresponding is as follows:
information. Additionally, a quality control test process is a. At the connection and communication
launched, in order to remove erroneous cylinders. This levels, it is selected standards for connec­
assembly line produces three different kinds of cylinders, tion, communication, and data format, in
one with a metal body and two with a plastic body. They order to deal with the standardization and
have differences in the size of the drilled holes between heterogeneity challenges. Each actor in the
the types of pistons. production process must have its cyber
As can be seen from the previous paragraph, this component, characterized as an intelligent
case study corresponds to a production line that has agent. In this case, the standards FIPA (2005)
some kind of automation, but with low autonomy. This for connection and communication can be
case study corresponds to a factory using Industry 3.0 useful, due that they are well defined and
concept. To transform this production line into the widely tested in industry. However, it could
Industry 4.0 context, we can proceed as follows: pose another challenge: ¿How to integrate
agents and cloud computing services? In
(1) Firstly, it is needed to transform all the devices, this case, Sanchez, Aguilar, and Exposito
like the drillers, the conveyor belts, the robot (2018b; (2018a)) have worked around this
arms, the manipulators into smart devices with issue, interconnecting the agents with ser­
capabilities of connectivity, communication, and vices deployed in the cloud, in order to
reasoning, among others. The actors in this pro­ extend their capabilities with cloud comput­
duction process can be represented as cyber ing services. Additionally, there are needed
components. A Multi-agent system that correctly some mechanisms to ensure that the data
characterizes each actor and can control the sen­ being shared among devices is protected
sors and effectors of the devices might be useful against unauthorized access.
for this purpose. b. The coordination, cooperation, and collabora­
(2) Secondly, the people involved in the production tion levels must be added, depending on the
line will be changed into devices that can accom­ production process characteristics. For the
plish their same functionality, entirety, or by using case study previously presented, it needs
some kind of HMI. It must reduce risk to people some coordination mechanisms, in order to
by keeping them away from the production line. allow:
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1037

i. The planning of the production strategy whole system. In this case, a CPS let services and
(self-planning). devices to interoperate in order to achieve production
ii. The execution of the plan (self-manage). goals, as well as to improve the effectiveness and
iii. The redefinition of the plan in case of fail­ efficiency of the entire industry.
ures (self-supervising and self-healing). IoT/IoE is used as an integration layer, which not
c. Mining techniques can be useful to deal with only allows the actors of the production process to
the complexity of planning and decision- communicate and interoperate but also to extend
making challenges regarding the coordination their capabilities by using services deployed through
process described above. Essentially, the the cloud computing paradigm (as is detailed in the
Mining techniques will allow creating knowl­ case study). In this sense, devices, people, data, and
edge-bases needed for smart decisions and services can connect and communicate using the
planning. For instance, the process mining standards provided by IoS and cloud computing.
techniques applied to the current production That characteristic is essential to promote autono­
process will be useful to get insights about mous processes for coordination, cooperation, and
how to coordinate the actors in the production collaboration, in order to drive the interoperability of
of the cylinders around a production plan that actors and allow them to achieve, intelligently and
defines how and when each actor will execute efficiently, collective and individual goals.
his tasks. Big data analytic techniques and, As can be seen, Industry 4.0 is fully integrated, so,
more precisely, machine learning algorithms System Integration is a crucial aspect in this context
are essential to predict the quality. They will because it allows integrating recent technologies (as is
help to check whether or not a cylinder is discussed in the case study) and putting them to work
going to fail the test before it enters the pro­ together in order to increase the autonomy of the pro­
duction line. In such a case, the coordination duction process. In the same sense, Modern Human-
and production plan must be redefined to Computer Interactions is very useful in Industry 4.0,
avoid failure. It will considerably reduce the because it allows people to be integrated into the man­
number of erroneous cylinders (Xu et al. 2017). ufacturing process transparently. For instance, in the
d. The negotiation challenge must be solved case study, people are separated from the production
by using the multi-agent system negotiation line, in order to reduce the occupational accident risks.
protocols (see (Calvaresi et al. 2020)). It will Technologies like wearables, AR, etc. allow people to
help to solve conflicts during the executions interoperate with other actors in a more intuitive way,
of the cyber component’s tasks. increasing cooperation, collaboration, and coordination
of those actors to improve production processes, redu­
cing waiting times, increasing security, and saving costs,
5. Conclusions
which are valuable characteristics in Industry 4.0. Finally,
Industry 4.0 is a concept still in development that arises Big Data Analysis is essential in Industry 4.0, because it
from the integration of technologies such as Artificial allows dealing with the heterogeneity of data and actors
Intelligence, smart factories, CPS, Cloud computing and by pre-processing large volume of data in looking for
IoS, IoT and IoE, Systems Integration, Modern Human- knowledge, identifying useful patterns, production
Computer Interactions, and Big Data Analysis, among models, among others.
others. In that sense, Smart Factory is an essential However, as is shown in the case study, a traditional
feature because it endows autonomy to the production industry cannot be turned into the Industry 4.0 concept
process, allowing it to self-configuring, self-supervising, only by integrating all the technologies described
self-healing, among others. In a Smart Factory, devices above. Industry 4.0 needs to endow autonomy to pro­
interoperate autonomously intending to achieve man­ duction processes. That means, not only autonomous
ufacturing goals, taking care of efficiency, and resource interoperability of actors, autonomous decision-
usage. Besides, CPS is another crucial technology used making, autonomous negotiation, etc., but also, self-
in Industry 4.0 to bring autonomy to the production configuring, self-healing, and self-supervising, among
process, due that CPS uses a smart control loop that other properties that bring autonomy to the produc­
allows adapting and improving the efficiency of the tion process. In that sense, autonomous coordination,
1038 M. SANCHEZ ET AL.

cooperation, and collaboration processes will be useful Systems: Results from a Systematic Literature Review.”
to let actors organizing, conjointly act, and efficiently 224–235. doi:10.5220/0006594802240235.
drive the production processes. Chang, V., and G. Wills. 2016. “A Model to Compare Cloud and
Non-Cloud Storage of Big Data.” Future Generation Computer
The future works are oriented to design and imple­
Systems 57 (April): 56–76. doi:10.1016/j.future.2015.10.003.
ment a framework that addresses the integration and Chen, K., S.-T. Hsu, J.-S. Jhang, and C.-S. Lin. 2017a. “Internet of
interoperability issues throughout the 5 C stack, in Things Security Appliance.” United States Patent Application
order to allow the production processes to self- 20170289176 Kind Code: A1, filed March 31, 2016, and
configuring, self-manage, self-healing, and self- issued October 5, 2017. http://www.freepatentsonline.com/
supervising. Specifically, the authors are going to y2017/0289176.html
Chen, X., P. Yang, T. Qiu, H. Yin, and J. Jianwei. 2017b. “IoE-MPP:
develop a solution for the coordination of actors
A Mobile Portal Platform for Internet of Everything.” Journal
related to the case study presented in section 4. of Intelligent & Fuzzy Systems 32 (4): 3069–3080. doi:10.3233/
Besides, our solution must be easily coupled to refer­ JIFS-169250.
ences architectures for Industry 4.0 like RAMI 4.0 Cheung, C. M. K., P.-Y. Chiu, and M. K. O. Lee. 2011. “Online
(Pisching et al. 2018; Platform Industry 4.0 2018) and Social Networks: Why Do Students Use Facebook?”
IIRA (Lin et al. 2015) in order to extend the standar­ Computers in Human Behavior, Social and Humanistic
Computing for the Knowledge Society, 27 (4): 1337–1343.
dized solution that already exists for Industry 4.0.
doi:10.1016/j.chb.2010.07.028.
Deloitte Consulting. 2017. “The Smart Factory.” Delloite
University Press. https://www.google.co.ve/url?sa=t&rct=
Disclosure statement j&q=&esrc=s&source=web&cd=3&ved=0ahUKEwjN3eWstv_
XAhWKOxQKHfk8AaIQFghHMAI&url=https%3A%2F%
No potential conflict of interest was reported by the authors. 2Fdupress.deloitte.com%2Fcontent%2Fdam%2Fdup-us-en
%2Farticles%2F4051_The-smart-factory%2FDUP_The-smart
-factory.pdf&usg=AOvVaw3XZS9-BuP0iQ9EL7EHPRgw
ORCID Drăgan, I., T. Selea, and F. Teodor-Florin. 2017. “Towards the
Integration of a HPC Build System in the Cloud Ecosystem.”
Manuel Sanchez http://orcid.org/0000-0002-3310-3394 In Complex, Intelligent, and Software Intensive Systems,
Ernesto Exposito http://orcid.org/0000-0002-3543-2909 916–925. Advances in Intelligent Systems and Computing.
Jose Aguilar http://orcid.org/0000-0003-4194-6882 Cham: Springer. doi:10.1007/978-3-319-61566-0_87.
Elattar, M., V. Wendt, and J. Jasperneite. 2017.
“Communications for Cyber-Physical Systems.” In Industrial
References Internet of Things, 347–372. Springer Series in Wireless
Technology. Cham: Springer. doi:10.1007/978-3-319-42559-
Aazam, M., and E.-N. Huh. 2016. “Fog Computing: The 7_13.
Cloud-IoT\/IoE Middleware Paradigm.” IEEE Potentials 35 Exposito, E. 2013. Advanced Transport Protocols: Designing the
(3): 40–44. doi:10.1109/MPOT.2015.2456213. Next Generation. London, UK: John Wiley & Sons.
Auger, A., E. Exposito, and E. Lochin. 2017. “Survey on Quality of Exposito, E., and C. Diop. 2014. Smart SOA Platforms in Cloud
Observation within Sensor Web Systems.” IET Wireless Sensor Computing Architectures. Networks And
Systems 7 (6): 163–177. doi:10.1049/iet-wss.2017.0008. Telecommunications. Hoboken, NJ, USA: Wiley-ISTE.
Bohuslava, J., J. Martin, and H. Igor. 2017. “TCP/IP Protocol doi:10.1002/9781118761489.
Utilisation in Process of Dynamic Control of Robotic Cell Ferrera, E., A. J. Rosaria Rossini, S. E. Baptista, G. G. Hovest,
According Industry 4.0 Concept.” In 2017 IEEE 15th M. Holgado, E. Lezak, et al. 2017. “Toward Industry 4.0:
International Symposium on Applied Machine Intelligence Efficient and Sustainable Manufacturing Leveraging
and Informatics (SAMI), 000217–22. Herl’any, Slovakia: IEEE. MAESTRI Total Efficiency Framework.” In Sustainable Design
doi:10.1109/SAMI.2017.7880306. and Manufacturing 2017, 624–633. Smart Innovation,
Büttner, S., H. Mucha, M. Funk, T. Kosch, M. Aehnelt, S. Robert, Systems and Technologies. Cham: Springer. doi:10.1007/
and R. Carsten. 2017. “The Design Space of Augmented and 978-3-319-57078-5_59.
Virtual Reality Applications for Assistive Environments in FIPA. 2005. “FIPA Available Specifications.” FIPA Specifications.
Manufacturing: A Visual Approach.” In Proceedings of the http://www.fipa.org/specifications/
10th International Conference on PErvasive Technologies Gökalp, E., U. Şener, and P. Erhan Eren. 2017. “Development of
Related to Assistive Environments, 433–440. PETRA ’17. an Assessment Model for Industry 4.0: Industry 4.0-MM.” In
New York, NY, USA: ACM. doi:10.1145/3056540.3076193. Software Process Improvement and Capability Determination,
Calvaresi, D., K. Appoggetti, L. Lustrissimi, M. Marinoni, 128–142. Communications in Computer and Information
P. Sernani, A. F. Dragoni, and M. Schumacher. 2020. “Multi- Science. Cham: Springer. doi:10.1007/978-3-319-67383-
Agent Systems’ Negotiation Protocols for Cyber-Physical 7_10.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1039

Goossens, J., and P. Richard. 2017. Handbook of Cyber-Physical 2017 IEEE International Conference on Communications (ICC),
Systems. Multiprocessor Real-Time Scheduling. http://hdl. 1–6. doi:10.1109/ICC.2017.7996801.
handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/ Kipper, G., and J. Rampolla. 2012. Augmented Reality: An
250127 Emerging Technologies Guide to AR. 1 ed. Amsterdam;
Grangel-González, I., L. Halilaj, G. Coskun, S. Auer, D. Collarana, Boston, MA: Syngress.
and M. Hoffmeister. 2016. “Towards a Semantic Lee, D., K.-H. Choi, and H. Kim. 2017. “Editorial: Smart Devices &
Administrative Shell for Industry 4.0 Components.” In 2016 Smart Spaces in Wireless Internet of Everything (Wireless-
IEEE Tenth International Conference on Semantic Computing ioe).” Wireless Personal Communications 94 (2): 145–147.
(ICSC), 230–237. doi:10.1109/ICSC.2016.58. doi:10.1007/s11277-017-4103-9.
Gupta, H., A. V. Dastjerdi, S. K. Ghosh, and R. Buyya. 2016. Lee, J., B. Bagheri, and H.-A. Kao. 2015. “A Cyber-Physical
“IFogSim: A Toolkit for Modeling and Simulation of Systems Architecture for Industry 4.0-Based Manufacturing
Resource Management Techniques in Internet of Things, Systems.” Manufacturing Letters 3 (Supplement C): 18–23.
Edge and Fog Computing Environments.” Software: doi:10.1016/j.mfglet.2014.12.001.
Practice & Experience 47 (9): 1275–1296. doi:10.1002/ Li, D., H. Tang, S. Wang, and C. Liu. 2017. “A Big Data Enabled
spe.2509. Load-Balancing Control for Smart Manufacturing of Industry
Hao, Y., and P. Helo. 2017. “The Role of Wearable Devices in 4.0.” Cluster Computing 20 (2): 1855–1864. doi:10.1007/
Meeting the Needs of Cloud Manufacturing: A Case Study.” s10586-017-0852-1.
Robotics and Computer-integrated Manufacturing, Special Li, X., L. Di, J. Wan, A. V. Vasilakos, C.-F. Lai, and S. Wang. 2017a.
Issue on Ubiquitous Manufacturing (UbiM), 45 (June): “A Review of Industrial Wireless Networks in the Context of
168–179. doi:10.1016/j.rcim.2015.10.001. Industry 4.0.” Wireless Networks 23 (1): 23–41. doi:10.1007/
Hofmann, E., and R. Marco. 2017. “Industry 4.0 And the Current s11276-015-1133-7.
Status as Well as Future Prospects on Logistics.” Computers Li, X., L. Di, J. Wan, A. V. Vasilakos, C.-F. Lai, and S. Wang. 2017b.
in Industry 89 (Supplement C): 23–34. doi:10.1016/j. “A Review of Industrial Wireless Networks in the Context of
compind.2017.04.002. Industry 4.0.” Wireless Networks 23 (1): 23–41. doi:10.1007/
Huang, Z., Y. Hang, Z. Peng, and Y. Feng. 2017. “Planning s11276-015-1133-7.
Community Energy System in the Industry 4.0 Era: Liao, Y., F. Deschamps, E. D. F. R. Loures, and L. F. P. Ramos.
Achievements, Challenges and a Potential Solution.” 2017. “Past, Present and Future of Industry 4.0 - a Systematic
Renewable and Sustainable Energy Reviews 78 (Supplement Literature Review and Research Agenda Proposal.”
C): 710–721. doi:10.1016/j.rser.2017.04.004. International Journal of Production Research 55 (12):
Huber, A., and A. Weiss. 2017. “Developing Human-Robot 3609–3629. doi:10.1080/00207543.2017.1308576.
Interaction for an Industry 4.0 Robot: How Industry Lin, S.-W., B. Miller, J. Durand, R. Joshi, P. Didier, C. Amine,
Workers Helped to Improve Remote-HRI to Physical-HRI.” R. Torenbeek, et al. 2015. “Industrial Internet Reference
In Proceedings of the Companion of the 2017 ACM/IEEE Architecture.” Tech. Rep 2017-01–25. Industrial Internet
International Conference on Human-Robot Interaction, Consortium (IIC). https://www.iiconsortium.org/pdf/SHI-
137–138. HRI ’17. New York, NY, USA: ACM. doi:10.1145/ WAN%20LIN_IIRA-v1%208-release-20170125.pdf
3029798.3038346. Liu, H., H. Ning, M. Qitao, Y. Zheng, L. T. Jing Zeng, R. H. Yang,
Ivanov, D., B. Sokolov, and M. Ivanova. 2016. “Schedule and M. Jianhua. 2017.“A Review of the Smart World.” Future
Coordination in Cyber-Physical Supply Networks Industry Generation Computer Systems, September. doi:10.1016/j.
4.0.” IFAC-PapersOnLine, 8th IFAC Conference on future.2017.09.010.
Manufacturing Modelling, Management and Control MIM Longo, F., L. Nicoletti, and A. Padovano. 2017. “Smart Operators
2016, 49 (12): 839–844. doi:10.1016/j.ifacol.2016.07.879. in Industry 4.0: A Human-Centered Approach to Enhance
Jazdi, N. 2014. “Cyber Physical Systems in the Context of Operators’ Capabilities and Competencies within the New
Industry 4.0.” In 2014 IEEE International Conference on Smart Factory Context.” Computers & Industrial Engineering
Automation, Quality and Testing, Robotics, 1–4. doi:10.1109/ 113 (Supplement C): 144–159. doi:10.1016/j.cie.2017.09.016.
AQTR.2014.6857843. Lu, Y. 2017. “Industry 4.0: A Survey on Technologies,
Jirkovský, V., M. Obitko, and M. Vladimír. 2017. “Understanding Applications and Open Research Issues.” Journal of
Data Heterogeneity in the Context of Cyber-Physical Industrial Information Integration 6 (Supplement C): 1–10.
Systems Integration.” IEEE Transactions on Industrial doi:10.1016/j.jii.2017.04.005.
Informatics 13 (2): 660–667. doi:10.1109/TII.2016.2596101. Martino, B. D., L. Kuan-Ching, L. T. Yang, and A. Esposito. 2018.
Kang, N., C. Zhao, L. Jingshan, and J. A. Horst. 2016. “Trends and Strategic Researches in Internet of Everything.”
“A Hierarchical Structure of Key Performance Indicators for In Internet of Everything, 1–12. Internet of Things. Singapore:
Operation Management and Continuous Improvement in Springer. doi:10.1007/978-981-10-5861-5_1.
Production Systems.” International Journal of Production McCann, J., and D. Bryson. 2009. Smart Clothes and Wearable
Research 54 (21): 6333–6350. doi:10.1080/ Technology. New York, NY, USA: Woodhead Publishing
00207543.2015.1136082. Limited and CRC Press LLC.
Khan, M., W. Xiaotong, X. Xiaolong, and W. Dou. 2017. “Big Data Mezghani, E., E. Expósito, and K. Drira. 2017a. “A Model-Driven
Challenges and Opportunities in the Hype of Industry 4.0.” In Methodology for the Design of Autonomic and Cognitive
1040 M. SANCHEZ ET AL.

IoT-Based Systems: Application to Healthcare.” IEEE Scalable Cyber-Physical Systems, 103 (Supplement C): 1–2.
Transactions on Emerging Topics in Computational doi:10.1016/j.jpdc.2017.01.025.
Intelligence 1 (3): 224–234. doi:10.1109/TETCI.2017.2699218. Richert, A., M. Shehadeh, S. Müller, S. Schröder, and S. Jeschke.
Mezghani, E., E. Expósito, and K. Drira. 2017b. “An Autonomic 2016. “Robotic Workmates – Hybrid Human-Robot-Teams in
Cognitive Pattern for Smart IoT-Based System the Industry 4.0.” In International Conference on e-Learning,
Manageability: Application to Comorbidity Management.” 127–31. Kuala Lumpur, Malaysia: Academic Conferences
https://hal.laas.fr/hal-01651945/document International Limited.
Mohamudally, N. 2017. Smartphones from an Applied Research Riel, A., and M. Flatscher. 2017. “A Design Process Approach to
Perspective. Rijeka, Croatia: InTech. Strategic Production Planning for Industry 4.0.” In Systems,
Molano, J. I. R., J. M. C. Lovelle, C. E. Montenegro, Software and Services Process Improvement, 323–333.
J. J. R. Granados, and R. G. Crespo. 2017. “Metamodel for Communications in Computer and Information Science.
Integration of Internet of Things, Social Networks, the Cloud Cham: Springer. doi:10.1007/978-3-319-64218-5_27.
and Industry 4.0.” Journal of Ambient Intelligence and Riggins, F., and T. Keskin. 2017. “Introduction to Internet of
Humanized Computing. February 1–15. doi:10.1007/s12652- Things: Providing Services Using Smart Devices, Wearables,
017-0469-5. and Quantified Self Minitrack.” In Proceedings of the 50th
Nelles, J., S. Kuz, A. Mertens, and C. Schlick. 2016. “Human- Hawaii International Conference on System Sciences.
Centered Design of Assistance Systems for Production doi:10.24251/HICSS.2017.166.
Planning and Control. The Role of the Human in Industry Román-Ibáñez, V., A. Jimeno-Morenilla, and F. A. Pujol-López.
4.0.” In IEEE International Conference on Industrial Technology 2018. “Distributed Monitoring of Heterogeneous Robotic
(ICIT), 2099–2104. Taipei, Taiwan: IEEE. doi:10.1109/ Cells. A Proposal for the Footwear Industry 4.0.”
ICIT.2016.7475093. International Journal of Computer Integrated Manufacturing
Obitko, M., and J. Václav. 2015. “Big Data Semantics in Industry 31 (12): 1205–1219. doi:10.1080/0951192X.2018.1529432.
4.0.” In Industrial Applications of Holonic and Multi-Agent Romero, D., T. Wuest, J. Stahre, and D. Gorecky. 2017. “Social
Systems, 217–229. Lecture Notes in Computer Science. Factory Architecture: Social Networking Services and
Cham: Springer. doi:10.1007/978-3-319-22867-9_19. Production Scenarios through the Social Internet of Things,
Orellana, F., and R. Torres. 2019. “From Legacy-Based Factories Services and People for the Social Operator 4.0.” In Advances
to Smart Factories Level 2 according to the Industry 4.0.” in Production Management Systems. The Path to Intelligent,
International Journal of Computer Integrated Manufacturing Collaborative and Sustainable Manufacturing, 265–273. IFIP
32 (4–5): 441–451. doi:10.1080/0951192X.2019.1609702. Advances in Information and Communication Technology.
Perera, C., A. Zaslavsky, P. Christen, and D. Georgakopoulos. Cham: Springer. doi:10.1007/978-3-319-66923-6_31.
2014. “Sensing as a Service Model for Smart Cities Supported Sanchez, M., J. Aguilar, and E. Exposito. 2018a. “Fog Computing
by Internet of Things.” Transactions on Emerging for the Integration of Agents and Web Services in an
Telecommunications Technologies 25 (1): 81–93. Autonomic Reflexive Middleware.” Service Oriented
doi:10.1002/ett.2704. Computing and Applications 1–15. doi:10.1007/s11761-018-
Pierdicca, R., E. Frontoni, R. Pollini, M. Trani, and L. Verdini. 2017. 0238-0.
“The Use of Augmented Reality Glasses for the Application Sánchez, M., J. Aguilar, and E. Exposito. 2018b. “Integración
in Industry 4.0.” In Augmented Reality, Virtual Reality, and SOA-MAS En Ambientes Inteligentes.” DYNA 85 (206):
Computer Graphics, Vol. 10324: 389–401. Lecture Notes in 268–282. doi:10.15446/dyna.v85n206.68671.
Computer Science. Cham: Springer. doi:10.1007/978-3-319- Santos, M. Y., J. O. E Sá, C. Costa, J. Galvão, C. Andrade,
60922-5_30. B. Martinho, F. V. Lima, and E. Costa. 2017. “A Big Data
Pisching, M. A., M. A. O. Pessoa, F. Junqueira, D. J. Dos Santos Analytics Architecture for Industry 4.0.” In Recent Advances
Filho, and P. E. Miyagi. 2018. “An Architecture Based on RAMI in Information Systems and Technologies, 175–184. Advances
4.0 To Discover Equipment to Process Operations Required in Intelligent Systems and Computing. Cham: Springer.
by Products.” Computers & Industrial Engineering 125 doi:10.1007/978-3-319-56538-5_19.
(November): 574–591. doi:10.1016/j.cie.2017.12.029. Schwab, K. 2016. La cuarta revolución industrial. Madrid,
Platform Industry 4.0. 2018. “Reference Architectural Model España: Penguin Random House Grupo Editorial.
Industrie 4.0 (RAMI4.0) - An Introduction.” Platform Sengupta, S., N. Gupta, and N. Vinayak Advisor. 2017. Firewall
Industry 4.0. https://www.plattform-i40.de/PI40/Redaktion/ for Internet of Things. New Dehli: Indraprastha Institute of
EN/Downloads/Publikation/rami40-an-introduction.html Information Technology. https://repository.iiitd.edu.in/
Preuveneers, D., and E. Ilie-Zudor. 2017. “The Intelligent xmlui/handle/123456789/587
Industry of the Future: A Survey on Emerging Trends, Shaikh, S. F., M. T. Ghoneim, G. T. Sevilla, J. M. Nassar,
Research Challenges and Opportunities in Industry 4.0.” A. M. Hussain, and M. M. Hussain. 2017. “Freeform
Journal of Ambient Intelligence and Smart Environments 9 Compliant CMOS Electronic Systems for Internet of
(3): 287–298. doi:10.3233/AIS-170432. Everything Applications.” IEEE Transactions on Electron
Qiu, M., S. Garg, R. Buyya, Y. Bei, and H. Shiyan. 2017. “Special Devices 64 (5): 1894–1905. doi:10.1109/TED.2016.2642340.
Issue on Scalable Cyber–Physical Systems.” Journal of Shila, D. M., W. Shen, Y. Cheng, X. Tian, and X. S. Shen. 2017.
Parallel and Distributed Computing, Special Issue on “AMCloud: Toward a Secure Autonomic Mobile Ad Hoc
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 1041

Cloud Computing System.” IEEE Wireless Communications 24 21st International Conference on Process Control (PC),
(2): 74–81. doi:10.1109/MWC.2016.1500119RP. 258–262. doi:10.1109/PC.2017.7976223.
Soto, J. A., C. F. Tavakolizadeh, and D. Gyulai. 2019. “An Online Vaquero, L. M., L. Rodero-Merino, J. Caceres, and M. Lindner.
Machine Learning Framework for Early Detection of Product 2008. “A Break in the Clouds: Towards A Cloud Definition.”
Failures in an Industry 4.0 Context.” International Journal of SIGCOMM Computer Communication Review 39 (1): 50–55.
Computer Integrated Manufacturing 32 (4–5): 452–465. doi:10.1145/1496091.1496100.
doi:10.1080/0951192X.2019.1571238. Vizcarrondo, J., J. Aguilar, E. Exposito, and A. Subias. 2012.
Strozzi, F., C. Colicchia, A. Creazza, and N. Carlo. 2017. “ARMISCOM: Autonomic Reflective Middleware for
“Literature Review on the ‘Smart Factory’ Concept Using Management Service Composition.” In 2012 Global
Bibliometric Tools.” International Journal of Production Information Infrastructure and Networking Symposium (GIIS),
Research 55 (22): 6572–6591. doi:10.1080/ 1–8. Choroni, Venezuela.
00207543.2017.1326643. Wan, J., S. Tang, Z. Shu, D. Li, S. Wang, M. Imran, and
Suri, K., A. Cuccuru, J. Cadavid, S. Gérard, W. Gaaloul, and A. V. Vasilakos. 2016. “Software-Defined Industrial Internet
S. Tata. 2017. “Model-Based Development of Modular of Things in the Context of Industry 4.0.” IEEE Sensors Journal
Complex Systems for Accomplishing System Integration 16 (20): 7373–7380. doi:10.1109/JSEN.2016.2565621.
for Industry 4.0.” In 5th International Conference on Model- Weyer, S., M. Schmitt, M. Ohmer, and D. Gorecky. 2015.
Driven Engineering and Software Development, 487–495. “Towards Industry 4.0 - Standardization as the Crucial
doi:10.5220/0006210504870495. Challenge for Highly Modular, Multi-Vendor Production
Syberfeldt, A., O. Danielsson, and P. Gustavsson. 2017. Systems.” IFAC (International Federation of Automatic
“Augmented Reality Smart Glasses in the Smart Factory: Control) Hosting by Elsevier Ltd 48 (3): 579–584.
Product Evaluation Guidelines and Review of Available doi:10.1016/j.ifacol.2015.06.143.
Products.” IEEE Access 5: 9118–9130. doi:10.1109/ Xu, Y. A. N., Y. A. N. M. I. N. G. Sun, J. I. A. F. U. Wan,
ACCESS.2017.2703952. X. I. A. O. L. O. N. G. Liu, and Z. H. I. T. I. N. G. Song. 2017.
Terán, J., J. Aguilar, and M. Cerrada. 2017. “Integration in “Industrial Big Data for Fault Diagnosis: Taxonomy, Review,
Industrial Automation Based on Multi-Agent Systems Using and Applications.” IEEE Access 5: 138–146. doi:10.1109/
Cultural Algorithms for Optimizing the Coordination ACCESS.2017.2731945.
Mechanisms.” Computers in Industry 91 (October): 11–23. Yang, L. T., B. D. Martino, and Q. Zhang. 2017. “Internet of
doi:10.1016/j.compind.2017.05.002. Everything.” Mobile Information Systems 2017: 1–3.
Vachálek, J., L. Bartalský, O. Rovný, D. Šišmišová, M. Morháč, doi:10.1155/2017/8035421.
and L. Milan. 2017. “The Digital Twin of an Industrial Zanero, S. 2017. “Cyber-Physical Systems.” Computer 50 (4):
Production Line within the Industry 4.0 Concept.” In 2017 14–16. doi:10.1109/MC.2017.105.