International Journal of Computer Integrated

Manufacturing

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

Data analytics in quality 4.0: literature review and

future research directions

Alexandros Bousdekis, Katerina Lepenioti, Dimitris Apostolou & Gregoris

Mentzas

To cite this article: Alexandros Bousdekis, Katerina Lepenioti, Dimitris Apostolou &
Gregoris Mentzas (2023) Data analytics in quality 4.0: literature review and future research
directions, International Journal of Computer Integrated Manufacturing, 36:5, 678-701, DOI:
10.1080/0951192X.2022.2128219

To link to this article: https://doi.org/10.1080/0951192X.2022.2128219

© 2022 The Author(s). Published by Informa

UK Limited, trading as Taylor & Francis
Group.

Published online: 28 Sep 2022.

Submit your article to this journal

Article views: 7170

View related articles

View Crossmark data

Citing articles: 20 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
2023, VOL. 36, NO. 5, 678–701
https://doi.org/10.1080/0951192X.2022.2128219

Data analytics in quality 4.0: literature review and future research directions
Alexandros Bousdekisa, Katerina Lepeniotia, Dimitris Apostolou a,b
and Gregoris Mentzasa
a
Information Management Unit (IMU), Institute of Communication and Computer Systems (ICCS), National Technical University of Athens
(NTUA), Athens, Greece; bDepartment of Informatics, University of Piraeus, Piraeus, Greece

ABSTRACT ARTICLE HISTORY

The quality level in manufacturing processes increasingly concerns manufacturing firms, as they Received 10 November 2021
respond to pressures such as increasing complexity and variety of products, more complex value Accepted 12 September 2022
chains and shortened time-to-market. Quality management is becoming increasingly challenging KEYWORDS
as model variety, and highly complex products harbour the danger of distributing defective machine learning; predictive
products in the market. Data analytics has started gathering the interest of quality researchers quality; Industry 4.0; data
and practitioners, who investigate approaches, algorithms, and methods for supporting the analytics
manufacturing quality procedures in the context of Industry 4.0. This trend is facilitated by the
wide expansion of sensory technology and the accelerated adoption of information systems by the
manufacturing firms. Since quality and process control has been identified as one of the major
challenges with a high potential of big data analytics, in this paper we investigated the manufac­
turing quality research field from a data analytics perspective. Specifically, we examined the
existing literature, we provided clarity to the Quality 4.0 research field, we synthesized the
literature review outcomes, and we identified the research gaps and challenges. On top of them,
we proposed directions for future research.

Introduction
System (QMS) is a part of a management system
Innovation, model variety, and highly complex pro­ regarding quality, and it comprises activities by
ducts addressing the needs and wishes of sophisti­ which the organization identifies its objectives and
cated customers harbour the danger of distributing determines the processes and resources required to
defective products in the market (Nalbach et al. achieve desired results. What we recognize as today’s
2018; Schmitt et al. 2020). In consequence of these quality profession began during the middle of
developments, the number of recalls continues to the second industrial revolution (Radziwill 2018). As
reach new levels, entailing massive image losses quality gained more and more importance over time,
(Deloitte 2016). As they respond to pressures such many quality concepts and methods have emerged.
as increasing complexity and variety of products, Quality models, approaches, and practices have
more complex value chains and shortened time-to- evolved from inspection to quality control, quality
market, manufacturing firms are increasingly con­ assurance, quality management, and business excel­
cerned with the quality level in manufacturing pro­ lence. Several models, frameworks, and tools have
cesses (Köksal, Batmaz, and Testik 2011; Psarommatis been developed to support organizations in mana­
et al. 2020a). ging and improving quality in all activity sectors.
ISO 9000:2015, Quality management systems, fun­ These include the Quality Tools and methods (such
damentals and vocabulary standard (ISO 2000) as 100% inspection and Statistical Quality Control),
defines quality management as management with the ISO 9001 Quality Management International
regard to quality; quality assurance as part of quality Standards, continual improvement methodologies
management focused on providing confidence that such as Lean, Six Sigma, or Lean Six Sigma, the teach­
quality requirements will be fulfilled and quality con­ ing of quality gurus such as Juran, Crosby, Deming, or
trol as part of quality management focused on ful­ Taguchi, and the business excellence models, namely,
filling quality requirements. A Quality Management the EFQM (Europe), the MBNA (EUA), or the Deming

CONTACT Alexandros Bousdekis [email protected] Information Management Unit, Institute of Communication and Computer Systems (ICCS),
National Technical University of Athens, 9, Iroon Polytechniou str, Athens 157 80, Greece
© 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 679

(Japan) models or awards (Fonseca and Domingues quality management body of knowledge, encompass­
2018; Köksal, Batmaz, and Testik 2011). ing models, systems, techniques, tools and extensive
Despite the significance of quality management application experience, can support the planning,
and quality assurance in modern manufacturing com­ implementation, and improvement of Industry 4.0
panies (Gunasekaran, Subramanian, and Ngai 2019) processes.
that affect the whole product lifecycle, i.e. the design, Key enablers of Quality 4.0 are data analytics, mod­
the manufacturing, and the service stages (Pal, els and algorithms for the analysis of data sets, for
Franciosa, and Ceglarek 2014), existing methods and example, industrial process, asset and product data
tools rely largely on the experience and subjective (Tsai et al. 2015). With data analytics, Quality 4.0 is
judgment of the people involved, which makes meant to be proactive: adverse effects of quality flaws
them time-consuming and unreliable (Nalbach et al. have to be prevented before they become relevant in
2018). Therefore, too often, quality engineers make the actual use of a product (Kupper et al. 2019;
their decisions by using only intuition and/or qualita­ Nalbach et al. 2018; Berger et al. 2018). Proactivity
tive assessments (Zonnenshain and Kenett 2020). enables to measure and predict the quality of systems
Most existing quality methods have not been adapted and products far in advance of those provided by
to cope with the data-intensive modern manufactur­ traditional preventive approaches (Köksal, Batmaz,
ing environment and thus, have lost their effective­ and Testik 2011; Deloitte 2016; Eger et al. 2018a;
ness (Bai et al. 2017; Psarommatis et al. 2020b). The Psarommatis et al. 2020b; Zonnenshain and Kenett
emergence of IoT and the increasing use of sensors in 2020; Bousdekis et al. 2018a; Bousdekis, Apostolou,
the shopfloor for monitoring the manufacturing pro­ and Mentzas 2020). Moreover, the increasing variety
cess and the machines has been providing huge and complexity of manufacturing processes as well as
amounts of data that can be utilized, among others, the low-quality, low-quantity and poor-suitability of
in the context of quality management. manufacturing data pose significant challenges to the
In the context of Industry 4.0, Quality 4.0 has defect’s identification and prediction (Chongwatpol
emerged as the combination of quality management 2015; Gittler et al. 2019). Data analytics, often enabled
and improvement models and approaches with tech­ by machine learning algorithms, has being among the
nology to foster critical competencies and factors for main emerging proactivity enablers (Chiarini 2020;
organizational success (Sader, Husti, and Daroczi Zonnenshain and Kenett 2020; Dhamija and Bag
2021). Similarly, Total Quality Management 4.0 refers 2020). While the recent state of research contains
to the ecosystem that supports the integration several literature reviews on general applications of
between technology, quality and people, which ML in manufacturing (Wuest et al. 2016; Lee, Shin, and
results from the adaptation of quality management Realff 2018b; Belhadi et al. 2019; Vater, Harscheidt,
to the technologies of I4.0, in the industrial scenario and Knoll 2019; Diez-Olivan et al. 2019), reviews with
(de Souza et al. 2021). While Industry 4.0 is more focus on quality-related applications are rarely found
technology-centric and quality is customer-centric (Schmitt et al. 2020). Quality and process control has
but using technology as an enabler, both approaches been identified as one of the major challenges with
aim for improved performance and results. Product a high potential of big data analytics (Belhadi et al.
and process quality is required to allow Industry 4.0 to 2019).
improve flexibility and productivity fully. Conversely, The objectives of this paper are (i) to investigate
intelligent sensors, automation, and big data can pro­ the existing literature regarding data analytics and
vide data for Quality Management Systems and busi­ machine learning in manufacturing quality opera­
ness excellence models and support Statistical tions; (ii) to provide clarity on the research field of
Process Control (SPC) or Six Sigma at the process Quality 4.0; (iii) to synthesize the literature review in
level (Fonseca, Amaral, and Oliveira 2021). Fonseca, order to identify the existing research challenges; and,
Amaral, and Oliveira (2021) point out that Quality 4.0, (iv) to outline directions for future research. The rest
by combining quality management with digitalization of the paper is organized as follows: Section 2 pre­
and technology, provide a management and process sents the methodology of the literature review. Its
dimension to the digital transformation technology three main steps are addressed in the subsequent
driver and support the successful I4.0 adoption. The sections. Section 3 frames the literature review by
680 A. BOUSDEKIS ET AL.

defining its scope and review protocol, by identifying ● Framing the literature review: This step, pre­
related literature reviews, and by presenting sented in Section 3, aims at posing the objectives
a bibliometric analysis. Section 4 analyses the and the research questions of the literature
reviewed research works by providing review. It consists of the scope definition
a classification, analysis, and synthesis of these (Section 3.1), the identification of related litera­
works. Section 5 discusses the resulting research ture reviews (Section 3.2), the review protocol
gaps and challenges, and outlines the directions for (Section 3.3) and a bibliometric analysis
future work. Section 6 concludes the paper. (Section 3.4).
● Review results: This step, presented in Section 4,
aims at analysing the reviewed papers in order to
Literature review methodology identify the current status on the development of
data analytics algorithms in quality manage­
In this Section, we outline the methodology of the
ment. It consists of the classification of the
literature review which is based on the methodology
reviewed papers (Section 4.1), their analysis in
of Tranfield, Denyer, and Smart (2003), largely used in
terms of the approaches and methods used for
literature reviews for data analytics and information
quality data analytics (Section 4.2), and their
systems (Nguyen et al. 2018; Barbosa et al. 2018;
synthesis (Section 4.3).
Duan, Edwards, and Dwivedi 2019; Koivisto and
● Research gaps, challenges, and future direc­
Hamari 2019). The methodology is presented in
tions: This step, presented in Section 5, sum­
Figure 1.
marizes and discusses the main conclusions of
The adopted methodology consists of three main
the literature review with a focus on the identi­
steps:
fied research gaps and challenges which form
the directions for future research.

Framing the literature review

In this Section, we present the frame of our literature
review. We define the scope, we identify related lit­
erature reviews, we present the review protocol, and
we present a bibliometric analysis.

Scope definition

Quality management has a systems perspective that

encompasses the two prominent views: product-
oriented and process-oriented quality (Psarommatis
et al. 2020a). Product-oriented quality studies the
defects on the actual parts and tries to find
a solution, while process-oriented quality studies the
defects of the manufacturing equipment, and based
on those can evaluate whether the manufactured
products are good or not. The latter lays within the
predictive maintenance concept, which has been
extensively explored in the literature in the frame of
the recent IoT advancements (Lindström et al. 2020;
Psarommatis et al. 2020); see e.g. (Bousdekis et al.
2018b; Carvalho et al. 2019; Zonta et al. 2020). On
the other hand, product-oriented quality is mainly
Figure 1. The methodology of the literature review. governed by manual knowledge-based processes or
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 681

by statistical quantitative techniques such as Table 1. Related literature reviews.

Statistical Process Control (SPC) and Theory of Reference Year Scope
Köksal, 2011 Data mining for quality improvement
Constraints (TOC) (Chongwatpol 2015; Belhadi et al. Batmaz,
2019). and Testik
(2011)
Psarommatis 2020 Zero defect manufacturing
et al.
Related literature reviews (2020a)
Zonnenshain 2020 Quality 4.0: quality and reliability engineering in
The first attempts of exploiting data analytics algo­ and Kenett Industry 4.0
(2020)
rithms involved data mining techniques (Köksal, Psarommatis 2020 Product Quality Improvement in Industry 4.0 and
et al. ZDM
Batmaz, and Testik 2011). The authors performed (2020b)
a review on data mining applications for manufactur­ Chiarini 2020 Relationships between Industry 4.0 and quality
(2020) management and TQM
ing quality improvement. Since then, the related lit­ Sony, 2020 Essential ingredients for the implementation of
erature has been evolved at a slow pace, while the Antony, Quality 4.0
and
focus has been on the use of ICT advancements from Douglas
a strategic point of view. It was only recently that the (2020)

need of framing quality in the data-rich context of

Industry 4.0 was arisen. Review protocol
Psarommatis et al. (2020a) provided a literature
review on Zero Defect Manufacturing (ZDM) and Table 2 presents the review protocol. We limited our
identified four distinctive strategies, i.e. detection, search space to include only journals, books and con­
repair, prediction, and prevention. Zonnenshain ference publications. We exclude any grey literature
and Kenett (2020) reviewed research trends on like white papers and blog posts because their quality
quality engineering and Industry 4.0, and proposed may vary and can affect the validity of our results. The
future research directions for quality and reliability keywords were used in various combinations among
engineering. Psarommatis et al. (2020b) performed them for the different scientific databases. The
a literature review on the implementation of Lean reviewed papers were selected according to whether
Manufacturing, Six Sigma, Lean Six Sigma, Total they contribute to the development and application
Quality Management, and Theory of Constraints of data analytics methods and algorithms in the man­
philosophies and compared them with ZDM philo­ ufacturing quality domain. In this sense, the resulting
sophy. Chiarini (2020) analysed the literature in number of papers does not include research works
terms of relationships between Industry 4.0, quality with qualitative and subjective quality tools that have
management and TQM in order to identify the not been adapted to the data-intensive modern man­
most relevant aspects referring to Quality 4.0. ufacturing environment. Therefore, we concluded to
Four categories of topics emerged, namely, creat­ 48 papers to be reviewed.
ing value within the company through quality (big)
data, analytics; developing Quality 4.0 skills; custo­
Bibliometric analysis
mer value co-creation; cyber – physical systems
and ERP for quality assurance and control. Sony, In this Section, we provide a bibliometric analysis of
Antony, and Douglas (2020) performed a literature the reviewed papers. Figure 2 depicts the amount of
review in order to reveal the key ingredients for the identified published papers on quality data analy­
the effective implementation of Quality 4.0. The tics per year. It is obvious that, during the last years,
authors concluded with the following ones: hand­ there is an increasing trend on the development and
ling big data, improving prescriptive analytics, implementation of data-driven methods and algo­
using Quality 4.0 for effective vertical, horizontal rithms for quality management.
and end-to-end integration, using Quality 4.0 for Moreover, Table 3 presents the journals including
strategic advantage, leadership, training, organiza­ related papers along with the publisher and the num­
tional culture, and top management support. ber of papers identified in each one. The journals
The aforementioned literature reviews are sum­ including more than one related paper are Journal
marized in Table 1. of Intelligent Manufacturing, International Journal of
682 A. BOUSDEKIS ET AL.

Table 2. Review protocol.

Item Description
Keywords Quality, quality control, quality assurance, quality management, quality improvement, artificial intelligence, big data, data analytics, descriptive
analytics, predictive analytics, prescriptive analytics, machine learning, neural network, Industry 4.0, Quality 4.0, smart manufacturing, zero
defect manufacturing
Inclusion Papers with data analytics methods for manufacturing quality
Exclusion Papers with qualitative and knowledge-based methods
Publishers Springer, Taylor & Francis, Elsevier, MDPI, Wiley, IEEE, Emerald, ACM
Time January 2012 to December 2020
period

Figure 2. The amount of the reviewed papers per year.

Production Research, and Expert systems with Review results

Applications. Table 4 presents the conferences includ­
In this Section, we present the results of our review.
ing related papers along with the proceedings pub­
We classify the reviewed papers, we discuss their
lisher and the number of papers identified in each
proposed approaches and methods, and we create
one. The conferences including more than one related
a taxonomy of methods that have been used in qual­
paper are CIRP Conference on Manufacturing Systems
ity data analytics.
(CMS), International Conference on Flexible
Automation and Intelligent Manufacturing (FAIM),
CIRP Global Web Conference, and IEEE international Classification of reviewed papers
conference on industrial informatics (INDIN). It should
We classified the reviewed papers in five dimensions:
be noted that very few papers are derived from qual­
type of contribution, manufacturing lifecycle stage,
ity-related journals, since they have strongly focused
level of intelligence, data & information sources, and
on extensions, improvements, and applications of tra­
industry. Each dimension is presented in the following
ditional and manual approaches, relying on human
sub-sections.
judgement. Data analytics algorithms are usually out
of the scope of these journals.
In order to further analyse the selected papers, we Type of contributions
used the VOSviewer tool which constructs and visua­ We classified the reviewed papers according to the
lizes bibliometric networks. Figure 3 depicts the most type of their contribution: conceptual framework, sys­
influential authors in the portfolio of the reviewed tem design and development, algorithm, and com­
papers, while Figure 4 depicts the main themes men­ parison of methods. Table 5 presents this
tioned throughout the articles. classification.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 683

Table 3. Journals including related papers.

Journal Publisher Number of papers
Journal of Intelligent Springer 5
Manufacturing
International Journal of Taylor & Francis 3
Production Research
Expert systems with Elsevier 2
Applications
Processes MDPI 1
Quality Engineering Elsevier 1
Advanced Engineering Elsevier 1
Informatics
Computers & Industrial Elsevier 1
Engineering
Engineering with Computers Springer 1
Sustainability MDPI 1
Sensors MDPI 1
Quality and Reliability Wiley 1
Engineering International
Advanced Intelligent Systems Wiley 1
Machines MDPI 1
Applied Sciences MDPI 1
Advances in Manufacturing Springer 1
IEEE Transactions on IEEE 1
Automation Science and
Engineering
Journal of Manufacturing Elsevier 1
Processes
Journal of Dispersion Science Taylor & Francis 1
and Technology
The International Journal of Springer 1
Advanced Manufacturing
Technology
CIRP Annals - Manufacturing Elsevier 1
Technology
Array Elsevier 1
Applied Soft Computing Elsevier 1
Journal of Power Sources Elsevier 1
Industrial Management & Data Emerald 1
Systems

Manufacturing lifecycle stage papers dealing with the quality control stage, three of
We classified the reviewed papers according to the them proposing a conceptual framework. Most of
manufacturing lifecycle stage which they refer to. The these papers use the same dataset. The literature
literature review focuses on the manufacturing phase review revealed that, despite the expansion of data
of product lifecycle modelling and not on the design analytics algorithms, that is facilitated by the increas­
and the service phases. To this end, we provide ing amounts of data, the quality control stage is still
a zoom-in to the manufacturing phase, which was governed by traditional and manual approaches, such
separated in three stages, as depicted in Figure 5: as Six Sigma and Lean Manufacturing. Even some
process configuration, in-process quality, and quality quantitative methodologies, such as Theory of
control. Table 6 presents this classification by also Constraints and Statistical Process Control have lim­
presenting the respective papers. Most of the itations in incorporating and analyzing data from
research works deal with in-process quality, i.e. during a variety of data sources in the dynamic and complex
the operation of the production line. Process config­ manufacturing environment.
uration mainly relies on manual processes that are
executed based on the expert knowledge due to the Level of intelligence
high complexity, variety and criticality of the decisions Data analytics can be categorized into three main
that need to be taken. Therefore, data analytics algo­ stages characterized by different levels of difficulty,
rithms in process configuration have not been used to value, and intelligence (Lepenioti et al. 2020): (i)
a wide extent. Finally, we found a limited amount of descriptive analytics, answering the questions ‘What
684 A. BOUSDEKIS ET AL.

Table 4. Conferences including related papers.

Conference Proceedings Number of papers
CIRP Conference on Elsevier 4
Manufacturing Systems
International Conference Elsevier 3
on Flexible Automation
and Intelligent
Manufacturing
CIRP Global Web Elsevier 2
Conference
IEEE International IEEE 2
Conference on
Industrial Informatics
(INDIN)
IFIP International Springer 1
Conference on
Advances in Production
Management Systems
(APMS)
Prognostics and System IEEE 1
Health Management
Conference
Conference on Robust Elsevier 1
Manufacturing
(RoMaC)
CIRP Conference on Life Elsevier 1
Cycle Engineering
IEEE International IEEE 1
Conference on Big Data
International Conference Springer 1
on Business
Information Systems

has happened?’, ‘Why did it happen?’, but also ‘What following sources: manufacturing sensors, environ­
is happening now?’ (mainly in a streaming context); mental sensors, product tracking technology, video
(ii) predictive analytics, answering the questions & image, enterprise & operational systems, and pro­
‘What will happen?’ and ‘Why will it happen?’ in the cess knowledge and specification. The classification is
future; (iii) prescriptive analytics, answering the ques­ presented in Table 8. Most of the research works
tions ‘What should I do?’ and ‘Why should I do it?’. utilize a combination of at least two types of data
We classified the reviewed papers according to the sources.
level of intelligence as derived from the level of data
analytics maturity, i.e. descriptive, predictive, and pre­ Industry
scriptive analytics. As depicted in Figure 6, each stage We classified the reviewed papers according to the
is a prerequisite for the next one in order to reach the industry to which they were applied. This classifica­
desired level of intelligence. It is pointed out that data tion is presented in Table 9. Steel industry gathers the
pre-processing is a prerequisite step for transforming larger number of quality applications. Automotive,
the raw data in a format capable of being further domestic appliances, and electronics industries have
processed by the data analytics algorithms (Level 0). also gathered a significant research interest.
Depending on the data format and structure, as well
as the implemented algorithms, the complexity of
Analysis of reviewed papers
data pre-processing varies Table 7 presents the clas­
sification of the reviewed papers according to the For each paper, we identified the level of intelligence
level of intelligence they reach. per manufacturing information lifecycle stage, as
shown in Table 10. Overall, the majority of research
Data and information sources works deal with in-process quality algorithms, sys­
We classified the reviewed papers according to the tems and approaches, where they rely on descriptive
data and information source. We identified the analytics (Level 1) and on predictive analytics
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 685

Figure 3. The most influential authors in the portfolio.

(Level 2). Some research works reach Level 3 (prescrip­ Descriptive analytics drew the most attention of
tive analytics), although only two of them (He et al. researchers. However, descriptive analytics
2017; Lindström et al. 2020) propose new algorithms; approaches rely to a large extend to process knowl­
the rest are conceptual frameworks. Process config­ edge, thus proposing ad-hoc and domain-specific
uration is mainly at Level 1, while we found 1 paper solutions, not easily extensible to different industries
(Kim and Ryu 2020) reaching Level 2. In Level 3, we and manufacturing processes. On the other hand,
found 2 conceptual research works (Psarommatis predictive analytics algorithms take advantage of
et al. 2020a; Zonnenshain and Kenett 2020). As the available data to a larger extent. Prescriptive ana­
already mentioned, quality management is still gov­ lytics is the least explored area.
erned by manual and statistical approaches, and it has
benefitted the least from the expansion of data ana­ Conceptual frameworks
lytics methods and technologies. We found only 2 It was only recently that the literature started to frame
research works at Level 2. Although the potential of quality management in the recent advancements of
data analytics to quality control processes has been Industry 4.0, and big data from a conceptual point of
recently identified in the literature (Psarommatis et al. view. Psarommatis and Kiritsis (2018) presented
2020a; Zonnenshain and Kenett 2020), the develop­ a conceptual architecture of a scheduling tool for
ment of related methods, algorithms and systems is at product-oriented ZDM which consists of four strate­
its early stages. gies: detection, repair, prevention, and prediction. The
Figure 7 depicts the data sources that are exploited first three strategies are covered to some extent by
in the reviewed papers per each level of intelligence. traditional quality improvement philosophies, such as
686 A. BOUSDEKIS ET AL.

Figure 4. The main keywords mentioned throughout the articles.

Lean Manufacturing, Six Sigma, Theory of Constraints, quality engineering and the development of predic­
and Total Quality Management (Psarommatis tive and prescriptive algorithms as significant
Giannakopoulos et al. 2020b). The prediction strategy enablers of Quality 4.0. Teucke et al. (2018) outlines
asks for advanced ICT systems and data analytics an approach for integrating sensor-based quality data
algorithms capable of early identifying and anticipat­ into supply chain event management. Franciosa et al.
ing defects in the product and the process (2020) proposed a digital twin framework for assem­
(Psarommatis Giannakopoulos et al. 2020b). bly systems with compliant parts fusing sensors with
According to the authors, there are numerous sche­ deep learning and CAE simulations in order to enable
duling tools available in the literature, but they are the ‘Closed-Loop In-Process (CLIP) quality improve­
‘machine’ oriented instead of ‘product’ oriented. ment. Bousdekis et al. (2021) proposed a framework
Psarommatis et al. (2020a) extended this approach for implementing quality analytics for decision aug­
by unifying the product-oriented and the process- mentation through optimized human-AI interaction
oriented ZDM. Zonnenshain and Kenett (2020) pro­ in quality control. All proposed conceptual frame­
posed a Quality 4.0 framework aiming at structuring works provide a useful tool for structuring challenges
the challenges and opportunities of quality and relia­ and opportunities of Industry 4.0 technologies in sup­
bility engineering in the Industry 4.0 era. The authors port of quality management and for depicting
consider the integration of reliability models to approaches addressing new and emerging quality
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 687

Table 5. Classification of research works per type of contribution.

Type of
contribution References
Conceptual Psarommatis and Kiritsis (2018); Psarommatis et al. (2020a); Zonnenshain and Kenett (2020); Teucke et al. (2018); Wang (2013); Franciosa
framework et al. (2020); Bousdekis et al. (2021)
System design Schmitt et al. (2020); Peres et al. (2018); Angione et al. (2019); Lee et al. (2018); Eger et al. (2018b); Luckow et al. (2018); Kim and Ryu
and (2020); Hamzeh et al. (2020); Nalbach et al. (2018)
development
Algorithm Liu et al. (2019); Chatterjee et al. (2019); Bustillo and Correa (2012); Lokrantz, Gustavsson, and Jirstrand (2018); He et al. (2017); Bai et al.
(2017); Lee et al. (2016); Escobar and Morales-Menendez (2018); Frumosu and Kulahci (2018); Jin, Zhang, and Gu (2020); Lindström
et al. (2020); Lee et al. (2014); Konrad, Lieber, and Deuse (2013); Lieber et al. (2013); Lieber et al. (2012); Hao et al. (2016); Paul (2016); Li
et al. (2012); Chamkalani, Chamkalani, and Mohammadi (2014); Liu et al. (2019b); Sun, Yang, and Wang (2017); Wuest, Irgens, and
Thoben (2014); Escobar, Morales-Menendez, and Macias (2020); Schreiber et al. (2019); Eger et al. (2020); Wang et al. (2019); Teti
(2015); Sanchez-Marquez et al. (2020); Haleem, Bustreo, and Del Bue (2021); Carbery, Woods, and Marshall (2018); Liu et al. (2020);
Maurya (2016); Pavlyshenko (2016); Scheibel, Mangler, and Rinderle-Ma (2021); Gashi et al. (2021)
Comparison of Jun, Chang, and Jun (2020); Bai et al. (2017); Oliff and Liu (2017); Bai et al. (2019); Schnell et al. (2019); Kim et al. (2012); Chongwatpol
methods (2015); Ge et al. (2021)

inspection tools, data analytics and knowledge

management technologies. Luckow et al. (2018)
evaluated architectures, models and deployment
challenges related to the use of deep learning
techniques in the automotive manufacturing qual­
ity and logistics focusing on computer vision pro­
blems. Hamzeh et al. (2020) developed a semi-
automatic welding system using a linear track
and a Gas Metal Arc Welding (GMAW) power
Figure 5. The manufacturing lifecycle stages. source as well as a mechanism for data acquisition,
integration, and visualization using heterogeneous
problems. Few, however, are capable of being oper­ sensor data. Nalbach et al. (2018) proposed
ationalised and hence their practical applicability is a machine learning-based system that links the
limited. quality assurance to the design phase in order to
support predictive quality. The system embeds sev­
System design and development eral algorithms, namely logistic regression, naive
In this Section, we discuss the research works dealing Bayes, random forests, multi-layer perceptron-style
with contributions to the quality information systems neural networks (MLPs) and AdaBoost based on
design and development. decision trees. System-based approaches are useful
Peres et al. (2018) proposed a data model primarily for systems engineers and architects
applicable to multi-stage ZDM manufacturing for because they provide practical support in design­
facilitating the reduction of defects, the identifica­ ing and developing holistic quality management
tion of their root causes and the elimination of systems, encompassing and combining various
propagation along the line. Angione et al. (2019) algorithms and methods. On the other hand, they
described the integration challenges of a platform have limited applicability in the shop floor, and
employing multi-agent systems, smart on-line they are not intended to be of use by operators.

Figure 6. The levels of intelligence according to data analytics maturity.

688 A. BOUSDEKIS ET AL.

Table 6. Classification of research works per manufacturing lifecycle stage.

Manufacturing lifecycle
stage References
Process configuration Psarommatis et al. (2020a); Zonnenshain and Kenett (2020); Eger et al. (2018b); Lee et al. (2014); Konrad, Lieber, and Deuse (2013);
Chamkalani, Chamkalani, and Mohammadi (2014); Kim and Ryu (2020); Schnell et al. (2019); Chongwatpol (2015)
In-process Psarommatis and Kiritsis (2018); Psarommatis et al. (2020a); Zonnenshain and Kenett (2020); Jun, Chang, and Jun (2020); Schmitt
et al. (2020); Liu et al. (2019); Chatterjee et al. (2019); Bai et al. (2018); Peres et al. (2018); Bustillo and Correa (2012); Lokrantz,
Gustavsson, and Jirstrand (2018); Oliff and Liu (2017); He et al. (2017); Bai et al. (2017); Lee et al. (2016); Angione et al. (2019);
Escobar and Morales-Menendez (2018); Lee et al. (2018); Frumosu and Kulahci (2018); Jin, Zhang, and Gu (2020); Bai et al. (2019);
Teucke et al. (2018); Lindström et al. (2020); Wang (2013); Lieber et al. (2013); Lieber et al. (2012); Hao et al. (2016); Paul (2016);
Liu et al. (2019b); Sun, Yang, and Wang (2017); Wuest, Irgens, and Thoben (2014); Li et al. (2012); Luckow et al. (2018); Franciosa
et al. (2020); Escobar, Morales-Menendez, and Macias (2020); Schreiber et al. (2019); Eger et al. (2020); Hamzeh et al. (2020);
Wang et al. (2019); Kim et al. (2012); Teti (2015); Haleem, Bustreo, and Del Bue (2021); Liu et al. (2020); Scheibel, Mangler, and
Rinderle-Ma (2021)
Quality control Psarommatis et al. (2020a); Zonnenshain and Kenett (2020); Nalbach et al. (2018); Sanchez-Marquez et al. (2020); Carbery, Woods,
and Marshall (2018); Maurya (2016); Ge et al. (2021); Pavlyshenko (2016); Gashi et al. (2021); Bousdekis et al. (2021)

Table 7. Classification of research works per level of intelligence.

Level of
intelligence References
Level 1 Liu et al. (2019); Peres et al. (2018); Lokrantz, Gustavsson, and Jirstrand (2018); Oliff and Liu (2017); Lee et al. (2016); Angione et al. (2019);
Eger et al. (2018b); Bai et al. (2019); Teucke et al. (2018); Lee et al. (2014); Hao et al. (2016); Chamkalani, Chamkalani, and Mohammadi
(2014); Sun, Yang, and Wang (2017); Wuest, Irgens, and Thoben (2014); Luckow et al. (2018); Schreiber et al. (2019); Eger et al. (2020);
Schnell et al. (2019); Kim et al. (2012); Teti (2015); Chongwatpol (2015); Li et al. (2012); Konrad, Lieber, and Deuse (2013); Escobar and
Morales-Menendez (2018); Lieber et al. (2013); Haleem, Bustreo, and Del Bue (2021); Maurya (2016); Pavlyshenko (2016); Scheibel,
Mangler, and Rinderle-Ma (2021)
Level 2 Jun, Chang, and Jun (2020); Schmitt et al. (2020); Chatterjee et al. (2019); Bai et al. (2017); Bustillo and Correa (2012); Bai et al. (2018); Lee
et al. (2018); Frumosu and Kulahci (2018); Jin, Zhang, and Gu (2020); Lieber et al. (2012); Paul (2016); Liu et al. (2019b); Escobar,
Morales-Menendez, and Macias (2020); Kim and Ryu (2020); Hamzeh et al. (2020); Wang et al. (2019); Nalbach et al. (2018); Sanchez-
Marquez et al. (2020); Carbery, Woods, and Marshall (2018); Liu et al. (2020); Ge et al. (2021); Gashi et al. (2021)
Level 3 Psarommatis and Kiritsis (2018); Psarommatis et al. (2020a); Zonnenshain and Kenett (2020); He et al. (2017); Lindström et al. (2020);
Wang (2013); Franciosa et al. (2020); Tambe and Kulkarni (2015); Bousdekis et al. (2021)

Table 8. Classification of research works per data and information source.

Data and information
source References
Manufacturing Sensors Teti (2015); Kim et al. (2012); Hamzeh et al. (2020); Sun, Yang, and Wang (2017); Wuest, Irgens, and Thoben (2014); Luckow et al.
(2018); Hao et al. (2016); Lieber et al. (2013); Eger et al. (2018b); Chatterjee et al. (2019); Jun, Chang, and Jun (2020); Bai et al.
(2018); Liu et al. (2019); Peres et al. (2018); Angione et al. (2019); Lindström et al. (2020); Wang (2013); Escobar, Morales-
Menendez, and Macias (2020); Schreiber et al. (2019); Kim and Ryu (2020); Wang et al. (2019); Liu et al. (2019b); Bustillo and
Correa (2012); Lieber et al. (2012); Lee et al. (2018); Escobar and Morales-Menendez (2018); Liu et al. (2020); Pavlyshenko (2016)
Environmental sensors Teucke et al. (2018); Eger et al. (2018b)
Product tracking Lee et al. (2014)
technology
Video and Image Hamzeh et al. (2020); Jin, Zhang, and Gu (2020); Schreiber et al. (2019); Liu et al. (2019b); Haleem, Bustreo, and Del Bue (2021)
Enterprise and Peres et al. (2018); Lokrantz, Gustavsson, and Jirstrand (2018); Oliff and Liu (2017); He et al. (2017); Angione et al. (2019); Lindström
operational systems et al. (2020); Wang (2013); Kim and Ryu (2020); Wang et al. (2019); Schnell et al. (2019); Nalbach et al. (2018); Chongwatpol
(2015); Lee et al. (2014); Lee et al. (2018); Schmitt et al. (2020); Lee et al. (2016); Sanchez-Marquez et al. (2020); Carbery, Woods,
and Marshall (2018); Liu et al. (2020); Maurya (2016); Ge et al. (2021); Pavlyshenko (2016); Scheibel, Mangler, and Rinderle-Ma
(2021); Gashi et al. (2021); Bousdekis et al. (2021)
Process knowledge and Paul (2016); Li et al. (2012); Chamkalani, Chamkalani, and Mohammadi (2014); Schreiber et al. (2019); Schnell et al. (2019); Eger
specifications et al. (2020); Franciosa et al. (2020); Konrad, Lieber, and Deuse (2013); Lee et al. (2014); Bai et al. (2019); Frumosu and Kulahci
(2018); Schmitt et al. (2020); Bai et al. (2017); Lee et al. (2016); Peres et al. (2018); Lokrantz, Gustavsson, and Jirstrand (2018); He
et al. (2017); Angione et al. (2019); Lindström et al. (2020); Wang (2013); Eger et al. (2020); Lee et al. (2014); Sanchez-Marquez
et al. (2020); Scheibel, Mangler, and Rinderle-Ma (2021)

Algorithms Eger et al. (2018b) developed a methodology and

Process configuration. Process configuration is a tool for the identification of correlations within
mainly addressed with descriptive analytics, while it a multi-stage production system. Schnell et al. (2019)
mainly relies on the use of enterprise data in combi­ compared and evaluated several data mining algo­
nation with expert and process knowledge. rithms, such as Generalized Linear Model (GLM),
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 689

Table 9. Classification of research works per industry.

Industry References
Process industry Jun, Chang, and Jun (2020); Chamkalani, Chamkalani, and Mohammadi (2014); Frumosu and Kulahci (2018)
Electronics Schmitt et al. (2020); Chongwatpol (2015); Wuest, Irgens, and Thoben (2014); Kim et al. (2012); Li et al. (2012)
manufacturing
Machining Bai et al. (2017); Angione et al. (2019); Teti (2015); Jin, Zhang, and Gu (2020)
Textile Lee et al. (2016); Lee et al. (2014); Nalbach et al. (2018); Haleem, Bustreo, and Del Bue (2021)
Domestic Angione et al. (2019); Oliff and Liu (2017); Carbery, Woods, and Marshall (2018); Liu et al. (2020); Maurya (2016); Ge et al. (2021);
appliances Pavlyshenko (2016); Bousdekis et al. (2021)
Automotive Escobar and Morales-Menendez (2018); Luckow et al. (2018); Angione et al. (2019); Teucke et al. (2018); Lindström et al. (2020); Franciosa
et al. (2020)
Steel industry Konrad, Lieber, and Deuse (2013); Lieber et al. (2013); Lieber et al. (2012); Paul (2016); Bustillo and Correa (2012); Lee et al. (2018); He et al.
(2017); Wang et al. (2019); Hao et al. (2016); Chatterjee et al. (2019); Gashi et al. (2021)
Battery Schnell et al. (2019); Escobar, Morales-Menendez, and Macias (2020)
manufacturing
Other Liu et al. (2019); Schreiber et al. (2019); Kim and Ryu (2020); Liu et al. (2019b); Sun, Yang, and Wang (2017); Eger et al. (2018b); Eger et al.
(2020); Sanchez-Marquez et al. (2020)
Numerical Zonnenshain and Kenett (2020); Bai et al. (2017); Lokrantz, Gustavsson, and Jirstrand (2018); Bai et al. (2019); Wang (2013); Peres et al.
example, (2018); Hamzeh et al. (2020); Scheibel, Mangler, and Rinderle-Ma (2021)
simulated
experiment

Table 10. Level of intelligence per manufacturing lifecycle stage.

Manufacturing
information
lifecycle Level 1 Level 2 Level 3
Process Eger et al. (2018b); Lee et al. (2014); Konrad, Kim and Ryu (2020) Psarommatis et al. (2020a); Zonnenshain and
configuration Lieber, and Deuse (2013); Chamkalani, Kenett (2020)
Chamkalani, and Mohammadi (2014);
Schnell et al. (2019); Chongwatpol (2015)
In-process Liu et al. (2019); Peres et al. (2018); Lokrantz, Jun, Chang, and Jun (2020); Schmitt et al. Psarommatis et al. (2020a); Zonnenshain and
Gustavsson, and Jirstrand (2018); Oliff and (2020); Chatterjee et al. (2019); Bai et al. Kenett (2020); Psarommatis and Kiritsis
Liu (2017); Lee et al. (2016); Angione et al. (2017); Bustillo and Correa (2012); Bai (2018); He et al. (2017); Lindström et al.
(2019); Escobar and Morales-Menendez et al. (2018); Lee et al. (2018); Frumosu (2020); Wang (2013); Franciosa et al.
(2018); Bai et al. (2019); Teucke et al. and Kulahci (2018); Jin, Zhang, and Gu (2020)
(2018); Hao et al. (2016); Sun, Yang, and (2020); Lieber et al. (2012); Paul (2016); Liu
Wang (2017); Wuest, Irgens, and Thoben et al. (2019b); Escobar, Morales-
(2014); Luckow et al. (2018); Schreiber Menendez, and Macias (2020); Hamzeh
et al. (2019); Eger et al. (2020); Kim et al. et al. (2020); Wang et al. (2019); Liu et al.
(2012); Teti (2015); Li et al. (2012); Lieber (2020)
et al. (2013); Haleem, Bustreo, and Del Bue
(2021); Scheibel, Mangler, and Rinderle-
Ma (2021)
Quality control Maurya (2016); Pavlyshenko (2016) Nalbach et al. (2018); Sanchez-Marquez et al. Psarommatis et al. (2020a); Zonnenshain and
(2020); Carbery, Woods, and Marshall Kenett (2020); Bousdekis et al. (2021)
(2018); Ge et al. (2021); Gashi et al. (2021)

Artificial Neural Networks (ANN), Support Vector Bai et al. (2019) compared three-dimension reduc­
Regression (SVR), Decision Trees (DT), Random tion techniques, i.e. principal component analysis
Forest (RF), and Gradient Boosted Trees (GBT) in (PCA), locally linear embedding (LLE), and isometric
order to identify process dependencies and key qual­ mapping (Isomap) and evaluated their effect on SVM
ity drivers in battery manufacturing. Chongwatpol performance for predictive quality modelling.
(2015) proposed three alternative models for diagnos­ Konrad, Lieber, and Deuse (2013) proposed the use
ing root causes of defects and variations by applying of data mining algorithms for gaining transparency
logistic regression, decision tree, and artificial neural on quality properties of intermediate products in
network (ANN) aiming at explaining the characteris­ process control and, in particular, in the steel indus­
tics of defects that have a great impact on manufac­ try domain. Chamkalani, Chamkalani, and
turing yield and the quality of products. Lee et al. Mohammadi (2014) developed a diagnostic model
(2014) presented an intelligent system, using fuzzy for asphaltene stability in the petroleum industry.
association rule mining with a recursive process The model is based upon Least-Square Support
mining algorithm, to find the relationships between Vector Machine (LSSVM) as well as a combination
production process parameters and product quality. of two search algorithms to deal with the problem
690 A. BOUSDEKIS ET AL.

Figure 7. The exploitation of data sources per level of intelligence.

of parameter selection: coupled simulated annealing and Jirstrand (2018) proposed a machine learning
(CSA) and the Nelder and Mead Simplex method. framework using Bayesian networks combined with
Their results outperform similar approaches using expert knowledge, in order to model the causal rela­
neural network and empirical models existing in tionships between manufacturing stages and to iden­
literature. tify the root causes of failures and quality deviations.
Existing literature covers adequately Level 1 in pro­ Hao et al. (2016) proposed a model to represent the
cess configuration, providing a plethora of algorithms impact of tool wear on quality degradation.
for describing and generating insights from quality- Kim et al. (2012) compared seven novelty detection
related data. However, we found only one paper methods and three different dimensionality reduction
reaching Level 2 in process configuration. Kim and methods for detecting faulty wafers in semiconductor
Ryu (2020) applied a Convolutional Neural Network manufacturing. Sun, Yang, and Wang (2017) proposed
(CNN) in order to derive predictions for quality man­ a method based on the particle swarm optimization
agement in the molding industry. Finally, we found and the kernel extreme learning machine in resistance
no research works on algorithms for process config­ spot welding to target the accurate and fast joint
uration reaching Level 3, i.e. implementing prescrip­ quality identification. Teti (2015) applied multi-
tive analytics. sensor signal processing for the extraction and selec­
tion of signal features for pattern recognition. Lieber
In-process. The literature is rich on in-process quality et al. (2013) implemented data pre-processing and
analytics algorithms, mostly descriptive analytics feature extraction and combined supervised and
algorithms. unsupervised learning methods to identify opera­
Liu et al. (2019) developed a real-time quality mon­ tional patterns, and quality-related features. Oliff and
itoring algorithm based on deep belief network (DBN) Liu (2017) proposed a methodology incorporating the
for quality spectra. Schreiber et al. (2019) proposed an rule-based learning algorithms C4.5 and RIPPER
approach for optical quality assurance using various (Repeated Incremental Pruning to Produce Error
machine learning algorithms. Lokrantz, Gustavsson, Reduction).
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 691

Wuest, Irgens, and Thoben (2014) proposed an measures in laser drilled holes quality. Bai et al.
approach based on clustering and supervised learn­ (2018) developed a framework with deep restricted
ing for coping with the complexity and high- Boltzmann machine and the stack autoencoder and
dimensionality of product state data. Li et al. (2012) compared it with a feed forward neural network with
proposed the use of SVR (Support Vector Machine for one hidden layer and a least squares support vector
Regression) for improving the cell vernier forecasting machine with no hidden layers. Lee et al. (2018)
model to enhance the production yield in the Color developed a Cyber-Physical Production System
Filter manufacturing process. Escobar and Morales- (CPPS) for quality prediction and operation control
Menendez (2018) proposed a pattern recognition in metal casting. Several machine learning algorithms
methodology supported by a hybrid feature elimina­ such as decision tree, random forest, artificial neural
tion algorithm and optimal classification threshold network, and support vector machine were used for
search algorithm for the detection of rare quality quality prediction.
events. Eger et al. (2020) presented an approach to Jin, Zhang, and Gu (2020) proposed a self-
compensate the dimensional deviations of an inner monitoring system based on real-time camera images
contour of a turbine shaft at an early stage in the and deep learning algorithms to classify the various
aerospace industry. extents of delamination in a printed part through
Haleem, Bustreo, and Del Bue (2021) proposed an additive manufacturing and to predict the onset of
online testing system for measurement of nep defects warping. Wang et al. (2019) proposed a generative
by using imaging and computer vision techniques. neural network model for automatically predicting
The developed system directly captures yarn images work-in-progress product quality. An autoencoding
on a spinning frame and uses Viola-Jones object neural network is trained using raw manufacturing
detection algorithm for real-time detection of nep process data and the extracted features are reformed
defects. Scheibel, Mangler, and Rinderle-Ma (2021) as time-series and are fed into a multi-layer percep­
proposed an approach to extract dimensioning infor­ tron. Schmitt et al. (2020) developed an integrated
mation from engineering drawings and to integrate solution for predictive model-based quality inspec­
this information into the production process to facil­ tion on the basis of recorded process parameters.
itate and optimize quality control. The extraction pro­ Liu et al. (2019b) explored the relationship between
cess is based on 2D clustering. the welding process and welded quality by develop­
There is also a considerable number of papers ing a multiple sensor fusion system with principal
reaching Level 2, i.e. developing and implementing component analysis and support vector machine
both descriptive and predictive analytics algorithms. Escobar, Morales-Menendez, and Macias (2020) pro­
Bai et al. (2017) proposed a deep neural network posed a big data-driven process monitoring aimed at
(DNN), consisting of a deep belief network (DBN) in rare quality event detection using the Support Vector
the bottom and a regression layer on the top in order Machine, Logistic Regression, Naive Bayes and
to overcome the challenges of the shallow architec­ k-Nearest Neighbors learning algorithms.
ture in product quality. Jun, Chang, and Jun (2020) Bustillo and Correa (2012) proposed a predictive
proposed semi-supervised learning, time-series ana­ model based on Bayesian Networks to optimize deep
lysis, and classification models within a framework for drilling operations under high-speed conditions for
predicting defects and improve yield in continuous- the manufacturing of steel components. Lieber et al.
flow manufacturing. In order to perform quality and (2012) investigated how data mining techniques and
yield prediction, they implemented and compared intelligent machine-to-machine telematics could be
Artificial Neural Networks (ANN), Logistic Regression, used to predict internal quality issues of intermediate
Decision Tree, Random Forest, Linear Discriminant products. Frumosu and Kulahci (2018) proposed an
Analysis (LDA), Gaussian Naive Bayes (GNB), approach to make use of latent structure – based
K-Nearest Neighbor (KNN), and Support Vector methods in the pursuit of better predictions. Paul
Classifier (SVC). (2016) collected Experimental Forming Limit Curve
Chatterjee et al. (2019) applied adaptive neuron- (FLC) and tensile properties of various steel grades
fuzzy inference system and multi-gene genetic pro­ from the literature and developed a predictive
gramming, in order to predict several performance model based on non-linear regression. Liu et al.
692 A. BOUSDEKIS ET AL.

(2020) proposed an end-to-end unified product qual­ optimising globally control corrective actions. They
ity prediction framework in order to capture temporal have not been operationalised yet, their reliability
interactions among processes in manufacturing and and efficacy in real-world conditions are largely
assembly processes. To do this, they developed untested, and hence their applicability in the shop
a bidirectional serial – parallel LSTM devised as an floor is, for the moment, limited.
instantiated model of temporal-interactive model.
The extensive use of heterogeneous data sources
on the shop floor has facilitated an increased level of Mapping of reviewed papers to lifecycle stages and
intelligence of in-process quality algorithms and pre­ analytics methods
scriptive analytics algorithms have just started to
emerge. However, these works are limited, while In this section, we map the reviewed papers to the
they utilize domain-specific optimization approaches respective manufacturing lifecycle stage (i.e. process
rather than machine learning techniques and algo­ configuration, in process, quality control) and cate­
rithms (e.g. Tambe and Kulkarni 2015; He et al. 2017) gory of methods (i.e. statistical correlation, regression
analysis, rule-based learning, neural networks, SVM,
Quality control. As already mentioned, quality con­ deep learning, clustering, probabilistic models). Not
trol is governed by manual and traditional quality surprisingly, many papers use combinations of meth­
procedures. Therefore, some research works investi­ ods to address quality challenges. Consequently,
gate their transformation to data-driven approaches. these references are associated with more than one
For example, Sanchez-Marquez et al. (2020) proposed category of methods. Moreover, based on the defined
a method to study quality management systems and taxonomy, for each paper we identified whether the
predict key performance indicators of balanced scor­ methods used address descriptive, predictive, or pre­
ecards. Another challenge has to do with the scarcity scriptive analytics. For each manufacturing lifecycle
of related datasets in the literature. The rest of the stage (i.e. process configuration, in-process, quality
identified papers use the Bosch dataset, a publicly control), we classified the descriptive, predictive, and
available but highly anonymized dataset. Carbery, prescriptive analytics research areas and the asso­
Woods, and Marshall (2018) used Bayesian networks ciated categories of methods that have been used in
for defect detection. Maurya (2016) performed anom­ the literature. Therefore, we classified the research
aly detection and binary classification by using areas and their associated categories of methods
Gradient Boosting Machine and Bayesian optimiza­ according to the data analytics stage that they
tion in order to identify rare defects. Ge et al. (2021) address.
compared Federated Support Vector Machine and Data pre-processing is a pre-requisite step for all
Federated Random Forest algorithms with centralized the data analytics algorithms so that the data are
learning techniques for product failure predictions. logically organized, structured and prepared for
Pavlyshenko (2016) combined the XGBoost tree- feeding into the algorithms. In some cases, this pro­
based classifier, the generalized linear model, and cess is straightforward; however, sometimes the nat­
a Bayesian approach for logistic regression. Gashi ure of data asks for more advanced signal processing
et al. (2021) proposed an approach to improve End and dimensionality reduction techniques. Table 11
of Line testing when condition monitoring data are presents such techniques as derived from the litera­
missing in order to predict low-quality products or ture review. Table 12 presents the reviewed works
products with a high probability of failure over time. assigned to the related category of methods and
To do this, they used a classification model by com­ manufacturing lifecycle stage. Figure 8 provides
paring linear discriminant analysis with shrinkage, a map showing the methods that are used for each
Random Forest, and lightGBM and they employed data analytics stage (i.e. descriptive, predictive, pre­
a multi-component system view for explaining the scriptive) and for each manufacturing lifecycle stage
prediction results. Although quality control is still (i.e. process configuration, in-process, quality con­
supported primarily by manual and traditional quality trol). For each point, it indicates the number of
procedures, data-driven approaches offer the poten­ papers that are associated with the aforementioned
tial of fusing a broader spectrum and data, and hence, dimensions.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 693

Table 11. Data pre-processing techniques identified in the lit­ having sophisticated requirements for expert knowl­
erature review. edge. On the other hand, quality management is still
Data pre-processing
techniques References governed by traditional and manual approaches, rely­
Principal Component Teti (2015); Liu et al. (2019b); Bai et al. ing on human judgement. Even some quantitative
Analysis (PCA) (2019) methodologies, such as Statistical Process Control
Locally Linear Embedding Bai et al. (2019)
(LLE) have limitations in incorporating and analyzing data
Isometric mapping (Isomap), Bai et al. (2019)
Stepwise linear regression Kim et al. (2012)
from a variety of data sources in the dynamic and
Stepwise 1-SVM Kim et al. (2012) complex manufacturing environment, while they can­
Synthetic Minority Wang et al. (2019)
Oversampling not adapt to the available data and the specific pro­
ReliefF Escobar and Morales-Menendez (2018) cesses in different industries.
l-1 regularization Escobar and Morales-Menendez (2018)
Wavelet Packet Transform Teti (2015) Despite the expansion of Industry 4.0, quality pro­
(WPT) cedures still do not adopt a data-driven approach
Synthetic Minority Over- Gashi et al. (2021)
sampling Technique that has the potential to facilitate the extraction of
(SMOTE) unrevealed insights even ahead of time, e.g. by pre­
Edited Nearest Neighbors Gashi et al. (2021)
(ENN) dicting the upcoming defects. Even common enter­
prise systems that have been adopted in the
previous years, such as ERP, MES, CMMS, end of line
Research gaps, challenges and future directions
quality management systems, etc., have not been
In this Section, we discuss the main research gaps that seen as a potential generator of quality insights
were identified in the literature based on the review that can lead to quality process optimization.
outcomes, their corresponding challenges, and the Future research can focus on the development of
related directions for future research. We identified data-driven frameworks and algorithms capable of
five main areas: (i) providing data-driven problem coping with the complexity and variability of process
solving; (ii) developing predictive and prescriptive configuration. In addition, there is the need for the
analytics algorithms; (iii) combining multiple data development of appropriate algorithms and meth­
sources; (iv) combining data and knowledge; and, (v) ods for transforming the procedures of the quality
providing augmented analytics capabilities. As shown control stage from manual ones to data-driven by
in Table 13, the column ‘Research Gaps’ summarizes exploiting various data sources that currently stay
the main findings of our literature review. They have untouched.
been identified based on the conclusions derived
from the aforementioned literature review. These
Developing predictive and prescriptive analytics
gaps lead to the research challenges and the future
algorithms
directions.
The current focus in the literature is on in-process
quality algorithms. Indeed, there is a large amount
Providing data-driven problem solving
of descriptive and predictive algorithms, while the
The increasing use of sensors in the shop-floor pro­ literature has started investigating the potential of
vides new capabilities for improving quality in the prescriptive analytics. On the other hand, process
production line and facilitating timely and proactive configuration is mainly addressed with descriptive
response to machine malfunctioning and product analytics algorithms, while quality control has not
defects. To this end, in-process quality algorithms exploited the opportunities of machine learning.
have taken advantage of the large amounts of real- However, the existing approaches rely to a large
time data in order to provide timely and reliable extent to process knowledge as well. On the other
insights during the actual operation. However, pro­ hand, existing predictive analytics algorithms take
cess configuration and even more, quality control advantage of the available data to a larger extent.
stages have not benefit from the technological Prescriptive analytics is the least explored area,
advancements of data analytics methods in the because of, among others, the scarcity of data-
frame of Industry 4.0. Process configuration is charac­ driven prescriptive analytics algorithms (Lepenioti
terized by a high complexity and variability, thus et al. 2020).
694 A. BOUSDEKIS ET AL.

Table 12. Categories of methods per manufacturing lifecycle stage.

Process configuration In-process Quality control
Statistical Eger et al. (2018b); Kim and Ryu (2020) Jun, Chang, and Jun (2020) Sanchez-Marquez
correlation et al. (2020)
Regression Schnell et al. (2019); Chongwatpol Escobar and Morales-Menendez (2018); Jun, Chang, and Jun (2020); Bai et al. Pavlyshenko
analysis (2015) (2019); Frumosu and Kulahci (2018); Hao et al. (2016); Schmitt et al. (2020); (2016);
Escobar, Morales-Menendez, and Macias (2020); Frumosu and Kulahci
(2018); Paul (2016)
Rule-based Lee et al. (2014); Schnell et al. (2019); Lee et al. (2016); Oliff and Liu (2017); Jun, Chang, and Jun (2020); Lee et al. Ge et al. (2021);
learning Chongwatpol (2015); Schnell et al. (2018); Lieber et al. (2013); Jun, Chang, and Jun (2020); Lee et al. (2018); Gashi et al.
(2019) Schmitt et al. (2020) (2021)
Neural Schnell et al. (2019) Sun, Yang, and Wang (2017); Teti (2015); Wang et al. (2019); Lieber et al. -
networks (2013); Lieber et al. (2012); Jun, Chang, and Jun (2020)
Support Schnell et al. (2019); Bai et al. (2019) Jun, Chang, and Jun (2020); Wuest, Irgens, and Thoben (2014); Li et al. (2012); Ge et al. (2021)
Vector Liu et al. (2019b); Escobar, Morales-Menendez, and Macias (2020); Schmitt
Machine et al. (2020); Lee et al. (2018)
Deep Kim and Ryu (2020) Liu et al. (2019); Bai et al. (2017); Schreiber et al. (2019); Jin, Zhang, and Gu -
learning (2020); Bai et al. (2017); Liu et al. (2020)
Clustering Maurya (2016) Kim et al. (2012); Lieber et al. (2012); Jun, Chang, and Jun (2020); Lieber et al. Gashi et al. (2021)
(2013); Escobar, Morales-Menendez, and Macias (2020); Haleem, Bustreo,
and Del Bue (2021); Scheibel, Mangler, and Rinderle-Ma (2021)
Probabilistic Maurya (2016) Lokrantz, Gustavsson, and Jirstrand (2018); Hao et al. (2016); Jun, Chang, and Carbery, Woods,
models Jun (2020); Kim et al. (2012); Schmitt et al. (2020); Escobar, Morales- and Marshall
Menendez, and Macias (2020); Bustillo and Correa (2012) (2018)

Predictive analytics in quality control moves Our literature review revealed that most of the
beyond the traditional quality management research works process data derived from sensors and
approaches by providing the capability to foresee from enterprise systems. Other data sources, such as
upcoming quality issues ahead of time and to take environmental sensors, cameras, and product tracking
mitigating actions. Further, prescriptive analytics have technology (e.g. RFID), are rarely exploited in the existing
the potential to maximize the quality goals and at the quality analytics algorithms. Towards this direction, Wei,
same time mitigate the likely risks by recommending Wu, and Terpenny (2020) developed a decision-level data
optimal sequences of actions by considering organi­ fusion approach that transforms low-dimensional deci­
zations quality objectives. sions (i.e. predictions) made based on individual sensor
data such as temperature and vibration to high-
dimensional decisions for quality control. Data fusion
Combining multiple data sources approaches have also been studies in the context of
Today’s manufacturing environment is rich on data quality characterization in metal additive manufacturing
sources providing different levels of information for (Grasso, Gallina, and Colosimo 2018), automating shop
different operations. For example, equipment- floor operations (Chen and Jin 2018) and manufacturing
installed sensors generate real-time measurements process monitoring (Kong et al. 2020).
for indicators of machine degradation; MES provide It should also be noted that existing quality analy­
data related to productivity and efficiency; quality tics algorithms require the embodiment of expert
management systems store data related to defects; knowledge and process specifications. This implies
ERP systems store data related to production man­ that the user can provide their knowledge and exper­
agement and supply chain management; CRM sys­ tise in a structured way, while a large volume of
tems gather customer-driven data on reviews, knowledge still remains tacit. Future research may
complaints and service defects. In addition to these, focus on the development of methodologies and
there are usually specific guidelines and specifications technologies for data and information fusion from
for executing certain tasks, which are either recorded heterogeneous sources.
in a semi-structured or unstructured format or are
based on the expert knowledge and are not recorded
Combining data and knowledge
in any system. However, there are data silos that
typically hold manufacturers back from greater effi­ Pure data-driven quality management approaches
ciencies and cost savings. have limitations on their practical implementation,
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 695

Figure 8. The taxonomy of the methods with respect to the data analytics stage.

especially when their outcomes are against the opera­ research may focus on a generalised framework for
tor’s knowledge and experience. Optimized human- fusing data-driven with knowledge-based methods
machine collaboration is enabled when data-driven for manufacturing applications such as quality man­
and knowledge-based systems interact in order to agement in order to provide non-intrusive decision
take into account both the engineered knowledge augmentation.
(‘voice of experts’) and the extracted knowledge
from data (‘voice of data’) by providing non-intrusive
decision augmentation. To tackle this issue, several Providing augmented analytics capabilities
research works incorporate domain knowledge at
design time and combine it with data analytics algo­ One barrier in the wider adoption of data analytics in
rithms. For example, hybrid approaches for fault quality management is that machine learning algo­
detection and health monitoring combine data- rithms require advanced analytical skills for their train­
driven analysis with knowledge-based models to ing, configuration and interpretation. Therefore,
overcome a lack of data and to increase fault detec­ transforming data into valuable insights requires an
tion accuracy (Tidriri et al. 2016; Wilhelm et al. 2021). automatic way by creating an optimized human-
Hybrid approaches have also been applied for machine collaboration in quality procedures. The
dynamic scheduling of shop floor operations (Ma aim of augmented analytics is to allow manufacturing
et al. 2022) as well as for enhancing predictive control companies to use machine learning to automatically
capabilities of industrial robots (Chee, Jiahao, and extract insights and visualize relevant findings from
Hsieh 2022) data without having to write algorithms or build com­
Existing hybrid approaches often lead to domain- plex models. They also aim at optimizing the use of
specific solutions, with low flexibility and scalability data for decision-making to augment human intelli­
putting barriers to their wider application. Future gence and contextual awareness (Gartner 2018).
696 A. BOUSDEKIS ET AL.

Table 13. Categories of methods per manufacturing lifecycle stage.

Section Research gaps Research challenges Future directions
5.1 Providing Process configuration and quality control Process configuration is characterized by Development of data-driven frameworks and
data-driven stages have not benefit from data- a high complexity and variability. algorithms for the process configuration and
problem driven approaches. Quality management is still governed by quality control stages.
solving traditional and manual approaches, relying
on human judgement.
5.2 Developing Predictive analytics has mainly been Existing approaches rely to a large extend to Development of predictive and prescriptive
predictive exploited for in-process functions and process knowledge as well. analytics algorithms in order to support
and not for process configuration and The training of predictive analytics quality engineers with reliable prescriptions
prescriptive quality control stages. algorithms is challenging to the scarcity of to mitigate upcoming quality issues based
analytics Prescriptive analytics is an unexplored recorded past events. on data-driven predictions.
algorithms area. Prescriptive analytics has been addressed Exploitation of machine learning models for
with domain-specific optimization generic and scalable solutions.
methods rather than machine learning.
5.3 Combining Μost of the research works deal with data Today’s manufacturing environment is rich on Development of methodologies and
multiple derived from sensors and enterprise data sources providing different levels of technologies for data and information fusion
data systems. Other data sources are rarely information for different operations from heterogeneous sources.
sources exploited. There are data silos that typically hold
Existing quality analytics algorithms manufacturers back from greater
usually require the embodiment of efficiencies and cost savings.
expert knowledge and process
specifications to a large extent.
5.4 Combining Existing literature has not investigated Pure data-driven approaches have limitations Coupling of data-driven with knowledge-based
data and the interaction between data-driven on their practical implementation, methods in order to augment decision
knowledge and knowledge-based methods. especially when their outcomes are against making based on both data and knowledge
There are domain-specific solutions, the operator’s experience. elicted from experts.
with low flexibility and scalability.
5.5 Providing Machine learning algorithms require Transform data into valuable insights in an In the context of augmented analytics, data
augmented advanced analytical skills for their automatic way by creating an optimized analytics can be enhanced with the use of
analytics training, configuration and human-machine collaboration in quality conversational interfaces through intelligent
capabilities interpretation. procedures. digital assistants allowing all users to easily
Supporting the operator to perform their interact with data and insights.
manual tasks with human augmentation Augmented analytics may use machine
technologies. learning to automatically extract insights
and visualize relevant findings from data
without having to write algorithms or build
complex models.

Augmented analytics may include natural language Conclusions

processing and conversational interfaces, allowing all
The quality level in manufacturing processes increas­
users to interact through spoken and written lan­
ingly concerns manufacturing firms, as they respond
guage (Prat 2019).
to pressures such as increasing complexity and variety
In the context of augmented analytics, data analy­ of products, more complex value chains and shor­
tics can be further enhanced with the use of conver­ tened time-to-market. Quality management is one of
sational interfaces as well as the exploitation of the the evergreen research areas in the modern century
human knowledge representation through intelligent that affects the whole product lifecycle, i.e. the
digital assistants allowing all users to easily interact design, the manufacturing, and the service stages.
with data and insights. Augmented analytics is an Our study focuses on quality management in the
emerging topic that is still at its infancy. The commu­ context of manufacturing. In the future, we plan to
nication with the user by voice pose new challenges extend the scope of analysis of quality management
to the development and execution of data analytics to encompass suppliers, partners, customers and con­
algorithms. Apart from visualization dashboards, the sumers and the overall production and distribution
data analytics outcomes should be structured in supply chains.
a way that can be translated to speech and be com­ Data analytics has started gathering the interest of
prehensible by the users. On the other way around, quality researchers and practitioners, who investigate
they should be able to take as input parameters that approaches, algorithms, and methods for supporting
are derived from the human speech. the manufacturing quality procedures in the context
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 697

of Industry 4.0. This trend is facilitated by the wide ORCID

expansion of sensory technology and the accelerated
Dimitris Apostolou http://orcid.org/0000-0002-5815-8033
adoption of information systems by the manufactur­
ing firms. Since quality and process control has been
identified as one of the major challenges with a high References
potential of big data analytics, in this paper we inves­
Angione, G., C. Cristalli, J. Barbosa, and P. Leitão. 2019.
tigated the manufacturing quality research field from
Integration Challenges for the Deployment of a
a data analytics perspective. Specifically, we exam­ Multi-Stage Zero-Defect Manufacturing Architecture. In
ined the existing literature, we provided clarity to 2019 IEEE 17th International Conference on Industrial
the Quality 4.0 research field, we synthesized the Informatics (INDIN), Helsinki, Finland, Vol. 1, 1615–1620. IEEE.
literature review outcomes, and we identified the Bai, Y., Z. Sun, J. Deng, L. Li, J. Long, and C. Li. 2017.
research gaps and challenges. On top of them, we “Manufacturing Quality Prediction Using Intelligent Learning
Approaches: A Comparative Study.” Sustainability 10 (1): 85.
proposed directions for future research.
Bai, Y., Z. Sun, J. Deng, L. Li, J. Long, and C. Li. 2018.
“Manufacturing Quality Prediction Using Intelligent
Learning Approaches: A Comparative Study.” Sustainability
Acknowledgements 10 (1): 85.
Bai, Y., Z. Sun, B. Zeng, J. Long, L. Li, J. V. de Oliveira, and C. Li.
This work is partly funded by the European Union’s Horizon
2019. “A Comparison of Dimension Reduction Techniques
2020 project COALA “COgnitive Assisted agile manufacturing
for Support Vector Machine Modeling of Multi-Parameter
for a LAbor force supported by trustworthy Artificial
Manufacturing Quality Prediction.” Journal of Intelligent
Intelligence” (Grant agreement No 957296). The work pre­
Manufacturing 30 (5): 2245–2256.
sented here reflects only the authors’ view and the European
Barbosa, M. W., A. D. L. C. Vicente, M. B. Ladeira, and
Commission is not responsible for any use that may be made of
M. P. V. D. Oliveira. 2018. “Managing Supply Chain
the information it contains.
Resources with Big Data Analytics: A Systematic Review.”
International Journal of Logistics Research and Applications
21 (3): 177–200.
Disclosure statement Belhadi, A., K. Zkik, A. Cherrafi, and M. Yusof. 2019.
“Understanding the Capabilities of Big Data Analytics for
No potential conflict of interest was reported by the author(s).
Manufacturing Process: Insights from Literature Review
and Multiple Case Study.” Computers & Industrial
Engineering 137: 106099.
Funding Berger, D., M. Zaiß, G. Lanza, J. Summa, M. Schwarz,
H. G. Herrmann, M. Pohl, F. Günther, and M. Stommel.
The work was supported by the Horizon 2020 Framework
2018. “Predictive Quality Control of Hybrid Metal-CFRP
Programme [957296].
Components Using Information Fusion.” Production
Engineering 12 (2): 161–172.
Bousdekis, A., D. Apostolou, and G. Mentzas. 2020. “Predictive
Notes on contributors Maintenance in the 4th Industrial Revolution: Benefits,
Alexandros Bousdekis is a senior research scientist at the Business Opportunities, and Managerial Implications.” IEEE
Information Management Unit of the Institute of Engineering Management Review 48 (1): 57–62.
Communication and Computer Systems in Greece. Bousdekis, A., B. Magoutas, D. Apostolou, and G. Mentzas.
2018b. “Review, Analysis and Synthesis of
Katerina Lepenioti is a researcher and a Ph.D. candidate at the Prognostic-Based Decision Support Methods for Condition
Information Management Unit of the Institute of Based Maintenance.” Journal of Intelligent Manufacturing
Communication and Computer Systems in Greece. 29 (6): 1303–1316.
Bousdekis, A., N. Papageorgiou, B. Magoutas, D. Apostolou, and
Dimitris Apostolou is associate professor in the Department of
G. Mentzas. 2018a. “Enabling Condition-Based Maintenance
Informatics at the University of Piraeus and a Senior Researcher
Decisions with Proactive Event-Driven Computing.”
at the Institute of Communication and Computer Systems in
Computers in Industry 100: 173–183.
Greece.
Bousdekis, A., S. Wellsandt, E. Bosani, K. Lepenioti,
Gregoris Mentzas is full Professor of Management Information D. Apostolou, K. Hribernik, and G. Mentzas. 2021,
Systems, School of Electrical and Computer Engineering, September. “Human-AI Collaboration in Quality Control
National Technical University of Athens and Director of the with Augmented Manufacturing Analytics.” In IFIP
Information Management Unit (IMU), a multidisciplinary International Conference on Advances in Production
research unit (http://imu.ntua.gr). Management Systems, 303–310. Cham: Springer.
698 A. BOUSDEKIS ET AL.

Bustillo, A., and M. Correa. 2012. “Using Artificial Intelligence to evolution, Challenges and Research Agenda.” International
Predict Surface Roughness in Deep Drilling of Steel Journal of Information Management 48: 63–71.
Components.” Journal of Intelligent Manufacturing 23 (5): Eger, F., D. Coupek, D. Caputo, M. Colledani, M. Penalva,
1893–1902. J. A. Ortiz, H. Freiberger, and G. Kollegger. 2018a. “Zero
Carbery, C. M., R. Woods, and A. H. Marshall. 2018. “A Bayesian Defect Manufacturing Strategies for Reduction of Scrap
Network Based Learning System for Modelling Faults in and Inspection Effort in Multi-Stage Production Systems.”
Large-Scale Manufacturing.” In 2018 IEEE International Procedia Cirp 67: 368–373.
Conference on Industrial Technology (ICIT), Lyon, France, Eger, F., C. Reiff, B. Brantl, M. Colledani, and A. Verl. 2018b.
1357–1362. IEEE. “Correlation Analysis Methods in Multi-Stage Production
Carvalho, T. P., F. A. Soares, R. Vita, R. D. P. Francisco, J. P. Basto, and Systems for Reaching Zero-Defect Manufacturing.” Procedia
S. G. Alcalá. 2019. “A Systematic Literature Review of Machine Cirp 72: 635–640.
Learning Methods Applied to Predictive Maintenance.” Eger, F., C. Reiff, P. Tempel, M. C. Magnanini, D. Caputo,
Computers & Industrial Engineering 137: 106024. A. Lechler, and A. Verl. 2020. “Reaching Zero-Defect
Chamkalani, A., R. Chamkalani, and A. H. Mohammadi. 2014. Manufacturing by Compensation of Dimensional
“Hybrid of Two Heuristic Optimizations with LSSVM to Deviations in the Manufacturing of Rotating Hollow Parts.”
Predict Refractive Index as Asphaltene Stability Identifier.” Procedia Manufacturing 51: 388–393.
Journal of Dispersion Science and Technology 35 (8): Escobar, C. A., and R. Morales-Menendez. 2018. “Machine
1041–1050. Learning Techniques for Quality Control in High
Chatterjee, S., S. S. Mahapatra, V. Bharadwaj, B. N. Upadhyay, Conformance Manufacturing Environment.” Advances in
and K. S. Bindra. 2019. “Prediction of Quality Characteristics Mechanical Engineering 10 (2): 1687814018755519.
of Laser Drilled Holes Using Artificial Intelligence Escobar, C. A., R. Morales-Menendez, and D. Macias. 2020.
Techniques.” Engineering with Computers 37: 1–24. “Process-Monitoring-For-Quality—a Machine Learning-
Chee, K. Y., T. Z. Jiahao, and M. A. Hsieh. 2022. KNODE-MPC: Based Modeling for Rare Event Detection.” Array 7: 100034.
A Knowledge-Based Data-Driven Predictive Control Fonseca, L., A. Amaral, and J. Oliveira. 2021. “Quality 4.0: The
Framework for Aerial Robots 7: 2819- 2826. IEEE: IEEE EFQM 2020 Model and Industry 4.0 Relationships and
Robotics and Automation Letters. Implications.” Sustainability 13 (6): 3107. doi:10.3390/
Chen, X., and R. Jin. 2018, August. Data Fusion Pipelines for su13063107.
Autonomous Smart Manufacturing. In 2018 IEEE 14th inter­ Fonseca, L. M., and J. P. Domingues. 2018. “The Best of Both
national conference on automation science and engineering Worlds? Use of Kaizen and Other Continuous Improvement
(CASE) Munich, Germany, 1203–1208. IEEE. Methodologies Within Portuguese ISO 9001 Certified
Chiarini, A. 2020. “Industry 4.0, Quality Management and TQM Organizations.” The TQM Journal 30 (4): 321–334.
World. A Systematic Literature Review and a Proposed Franciosa, P., M. Sokolov, S. Sinha, T. Sun, and D. Ceglarek. 2020.
Agenda for Further Research.” The TQM Journal 32 (4): “Deep Learning Enhanced Digital Twin for Closed-Loop
603–616. doi:10.1108/TQM-04-2020-0082. In-Process Quality Improvement.” CIRP Annals 69 (1):
Chongwatpol, J. 2015. Prognostic Analysis of Defects in 369–372.
Manufacturing. United Kingdom: Industrial Management & Frumosu, F. D., and M. Kulahci. 2018. “Big Data Analytics Using
Data Systems. Semi-supervised Learning Methods.” Quality and Reliability
Deloitte. 2016. Predictive Quality Analytics: Key Factor on the Engineering International 34 (7): 1413–1423.
Way to Quality Leadership. Accessed 3 January 2021. https:// Gartner. 2018. Augmented Analytics is the Future of Data and
www2.deloitte.com/content/dam/Deloitte/de/Documents/ Analytics. Accessed 3 January 2021. https://www.gartner.
manufacturing/DELO_Predictive%20Quality%20Analytics_ com/en/documents/3990724/summary-translation-
ks3.pdf. augmented-analytics-is-the-future-of
de Souza, F. F., A. Corsi, R. N. Pagani, G. Balbinotti, and Gashi, M., P. Ofner, H. Ennsbrunner, and S. Thalmann. 2021.
J. L. Kovaleski. 2021. “Total Quality Management 4.0: “Dealing with Missing Usage Data in Defect Prediction:
Adapting Quality Management to Industry 4.0.” The TQM A Case Study of a Welding Supplier.” Computers in Industry
Journal 34 (4): 749–769. 132: 103505.
Dhamija, P., and S. Bag. 2020. “Role of Artificial Intelligence in Ge, N., G. Li, L. Zhang, and Y. Liu. 2021. “Failure Prediction in
Operations Environment: A Review and Bibliometric Production Line Based on Federated Learning: An Empirical
Analysis.” The TQM Journal 32 (4): 869–896. doi:10.1108/ Study.” Journal of Intelligent Manufacturing 1: 1572–8145.
TQM-10-2019-0243. Gittler, T., E. Relea, D. Corti, G. Corani, L. Weiss, D. Cannizzaro,
Diez-Olivan, A., J. Del Ser, D. Galar, and B. Sierra. 2019. “Data and K. Wegener. 2019. “Towards Predictive Quality
Fusion and Machine Learning for Industrial Prognosis: Management in Assembly Systems with Low Quality Low
Trends and Perspectives Towards Industry 4.0.” Information Quantity Data–a Methodological Approach.” Procedia Cirp
Fusion 50: 92–111. 79: 125–130.
Duan, Y., J. S. Edwards, and Y. K. Dwivedi. 2019. “Artificial Grasso, M., F. Gallina, and B. M. Colosimo. 2018. “Data Fusion
Intelligence for Decision Making in the Era of Big Data– Methods for Statistical Process Monitoring and Quality
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 699

Characterization in Metal Additive Manufacturing.” Procedia Kupper, D., C. Knizek, D. Ryeson, and J. Noecker (2019), “Quality
Cirp 75: 103–107. 4.0 Takes More Than Technology”. Accessed 3 January 2021.
Gunasekaran, A., N. Subramanian, and W. T. E. Ngai. 2019. https://www.bcg.com/publications/2019/quality-4.0-takes-
“Quality Management in the 21st Century Enterprises: more-than-technology.aspx
Research Pathway Towards Industry 4.0.” International Lee, C. K. H., K. L. Choy, G. T. Ho, and C. H. Lam. 2016. “A Slippery
Journal of Production Economics 207: 125–129. Genetic Algorithm-Based Process Mining System for
Haleem, N., M. Bustreo, and A. Del Bue. 2021. “A Computer Achieving Better Quality Assurance in the Garment
Vision Based Online Quality Control System for Textile Industry.” Expert Systems with Applications 46: 236–248.
Yarns.” Computers in Industry 133: 103550. Lee, C. K. H., G. T. S. Ho, K. L. Choy, and G. K. H. Pang. 2014. “A
Hamzeh, R., L. Thomas, J. Polzer, X. W. Xu, and H. Heinzel. 2020. RFID-Based Recursive Process Mining System for Quality
“A Sensor Based Monitoring System for Real-Time Quality Assurance in the Garment Industry.” International Journal
Control: Semi-Automatic Arc Welding Case Study.” Procedia of Production Research 52 (14): 4216–4238.
Manufacturing 51: 201–206. Lee, J., S. D. Noh, H. J. Kim, and Y. S. Kang. 2018.
Hao, L., L. Bian, N. Gebraeel, and J. Shi. 2016. “Residual Life “Implementation of Cyber-Physical Production Systems for
Prediction of Multistage Manufacturing Processes with Quality Prediction and Operation Control in Metal Casting.”
Interaction Between Tool Wear and Product Quality Sensors 18 (5): 1428.
Degradation.” IEEE Transactions on Automation Science and Lee, J. H., J. Shin, and M. J. Realff. 2018b. “Machine Learning:
Engineering 14 (2): 1211–1224. Overview of the Recent Progresses and Implications for the
He, Y., C. Gu, Z. Chen, and X. Han. 2017. “Integrated Predictive Process Systems Engineering Field.” Computers & Chemical
Maintenance Strategy for Manufacturing Systems by Engineering 114: 111–121.
Combining Quality Control and Mission Reliability Lepenioti, K., A. Bousdekis, D. Apostolou, and G. Mentzas. 2020.
Analysis.” International Journal of Production Research “Prescriptive Analytics: Literature Review and Research
55 (19): 5841–5862. Challenges.” International Journal of Information
ISO 9000, B. S. E. N. 2000. Quality Management Systems: Management 50: 57–70.
Fundamentals and Vocabulary. London: British Standards Li, D. C., W. C. Chen, C. W. Liu, and Y. S. Lin. 2012. “A Non-Linear
Institution. Quality Improvement Model Using SVR for Manufacturing
Jin, Z., Z. Zhang, and G. X. Gu. 2020. “Automated Real-time TFT-LCDs.” Journal of Intelligent Manufacturing 23 (3):
Detection and Prediction of Interlayer Imperfections in 835–844.
Additive Manufacturing Processes Using Artificial Lieber, D., B. Konrad, J. Deuse, M. Stolpe, and K. Morik.
Intelligence.” Advanced Intelligent Systems 2 (1): 1900130. 2012. “Sustainable Interlinked Manufacturing Processes
Jun, J. H., T. W. Chang, and S. Jun. 2020. “Quality Prediction and Through Real-Time Quality Prediction, Dornfeld, David,
Yield Improvement in Process Manufacturing Based on Data Linke, Barbara.” In Leveraging Technology for
Analytics.” Processes 8 (9): 1068. a Sustainable World, 393–398. Berlin, Heidelberg:
Kim, D., P. Kang, S. Cho, H. J. Lee, and S. Doh. 2012. “Machine Springer.
Learning-Based Novelty Detection for Faulty Wafer Lieber, D., M. Stolpe, B. Konrad, J. Deuse, and K. Morik. 2013.
Detection in Semiconductor Manufacturing.” Expert “Quality Prediction in Interlinked Manufacturing Processes
Systems with Applications 39 (4): 4075–4083. Based on Supervised & Unsupervised Machine Learning.”
Kim, S., and K. Ryu. 2020. “Intelligent Process Quality Procedia Cirp 7: 193–198.
Management for Supporting Collaboration of Mold Lindström, J., P. Kyösti, W. Birk, and E. Lejon. 2020. “An Initial
Manufacturing SMEs.” Procedia Manufacturing 51: 381–387. Model for Zero Defect Manufacturing.” Applied Sciences
Koivisto, J., and J. Hamari. 2019. “The Rise of Motivational 10 (13): 4570.
Information Systems: A Review of Gamification Research.” Liu, G., X. Gao, D. You, and N. Zhang. 2019b. “Prediction of High
International Journal of Information Management 45: 191–210. Power Laser Welding Status Based on PCA and SVM
Köksal, G., İ. Batmaz, and M. C. Testik. 2011. “A Review of Data Classification of Multiple Sensors.” Journal of Intelligent
Mining Applications for Quality Improvement in Manufacturing 30 (2): 821–832.
Manufacturing Industry.” Expert Systems with Applications Liu, Z., D. Zhang, W. Jia, X. Lin, and H. Liu. 2020. “An Adversarial
38 (10): 13448–13467. Bidirectional Serial–parallel LSTM-Based QTD Framework for
Kong, L., X. Peng, Y. Chen, P. Wang, and M. Xu. 2020. “Multi- Product Quality Prediction.” Journal of Intelligent
Sensor Measurement and Data Fusion Technology for Manufacturing 31 (6): 1511–1529.
Manufacturing Process Monitoring: A Literature Review.” Liu, Y., H. Zhou, F. Tsung, and S. Zhang. 2019. “Real-Time
International Journal of Extreme Manufacturing 2 (2): 022001. Quality Monitoring and Diagnosis for Manufacturing
Konrad, B., D. Lieber, and J. Deuse. 2013. “Striving for Zero Process Profiles Based on Deep Belief Networks.”
Defect Production: Intelligent Manufacturing Control Computers & Industrial Engineering 136: 494–503.
Through Data Mining in Continuous Rolling Mill Processes, Lokrantz, A., E. Gustavsson, and M. Jirstrand. 2018. “Root Cause
Windt, Katja.” In Robust Manufacturing Control, 215–229. Analysis of Failures and Quality Deviations in Manufacturing
Berlin, Heidelberg: Springer. Using Machine Learning.” Procedia Cirp 72: 1057–1062.
700 A. BOUSDEKIS ET AL.

Luckow, A., K. Kennedy, M. Ziolkowski, E. Djerekarov, M. Cook, Radziwill, N. M. 2018. ”Quality 4.0: Let’s Get Digital-The Many
E. Duffy, M. Schleiss, B. Vorster, E. Weill, A. Kulshrestha, and Ways the Fourth Industrial Revolution is Reshaping the Way
M. C. Smith. 2018. “Artificial Intelligence and Deep Learning We Think About Quality.“ arxiv 1: 24–29. (arXiv preprint
Applications for Automotive Manufacturing.” In 2018 IEEE arXiv:1810.07829).
International Conference on Big Data (Big Data), Seattle, Sader, S., I. Husti, and M. Daroczi. 2021. “A Review of Quality 4.0:
WA, USA, 3144–3152. IEEE. Definitions, Features, Technologies, Applications, and
Ma, Y., S. Li, F. Qiao, X. Lu, and J. Liu. 2022. “A Data-Driven Challenges.” Total Quality Management & Business
Scheduling Knowledge Management Method for Smart Excellence 1: 1164–1182.
Shop Floor.” International Journal of Computer Integrated Sanchez-Marquez, R., J. M. A. Guillem, E. Vicens-Salort, and
Manufacturing, 1-14: 780–793. J. J. Vivas. 2020. “Diagnosis of Quality Management
Maurya, A. 2016. Bayesian Optimization for Predicting Rare Systems Using Data Analytics–a Case Study in the
Internal Failures in Manufacturing Processes. In 2016 IEEE Manufacturing Sector.” Computers in Industry 115: 103183.
international conference on big data (big data), Washington Scheibel, B., J. Mangler, and S. Rinderle-Ma. 2021. “Extraction of
D.C., USA, 2036–2045. IEEE. Dimension Requirements from Engineering Drawings for
Nalbach, O., C. Linn, M. Derouet, and D. Werth. 2018. Supporting Quality Control in Production Processes.”
“Predictive Quality: Towards a New Understanding of Computers in Industry 129: 103442.
Quality Assurance Using Machine Learning Tools.” In Schmitt, J., J. Bönig, T. Borggräfe, G. Beitinger, and J. Deuse.
International Conference on Business Information Systems, 2020. “Predictive Model-Based Quality Inspection Using
30–42. Cham: Springer. Machine Learning and Edge Cloud Computing.” Advanced
Nguyen, T., Z. H. O. U. Li, V. Spiegler, P. Ieromonachou, and Engineering Informatics 45: 101101.
Y. Lin. 2018. “Big Data Analytics in Supply Chain Schnell, J., C. Nentwich, F. Endres, A. Kollenda, F. Distel,
Management: A State-Of-The-Art Literature Review.” T. Knoche, and G. Reinhart. 2019. “Data Mining in
Computers & Operations Research 98: 254–264. Lithium-Ion Battery Cell Production.” Journal of Power
Oliff, H., and Y. Liu. 2017. “Towards Industry 4.0 Utilizing Sources 413: 360–366.
Data-Mining Techniques: A Case Study on Quality Schreiber, M., J. Klöber-Koch, J. Bömelburg-Zacharias,
Improvement.” Procedia Cirp 63: 167–172. S. Braunreuther, and G. Reinhart. 2019. “Automated Quality
Pal, A., P. Franciosa, and D. Ceglarek. 2014. “Root Cause Analysis Assurance as an Intelligent Cloud Service Using Machine
of Product Service Failures in Design-A Closed-Loop Learning.” Procedia Cirp 86: 185–191.
Lifecycle Modelling Approach.” Procedia Cirp 21: 165–170. Sony, M., J. Antony, and J. A. Douglas. 2020. “Essential
Paul, S. K. 2016. “Prediction of Complete Forming Limit Ingredients for the Implementation of Quality 4.0.” The
Diagram from Tensile Properties of Various Steel Sheets by TQM Journal 32 (4): 779–793. doi:10.1108/TQM-12-2019-
a Nonlinear Regression Based Approach.” Journal of 0275.
Manufacturing Processes 23: 192–200. Sun, H., J. Yang, and L. Wang. 2017. “Resistance Spot Welding
Pavlyshenko, B. 2016. “Machine Learning, Linear and Bayesian Quality Identification with Particle Swarm Optimization and
Models for Logistic Regression in Failure Detection a Kernel Extreme Learning Machine Model.” International
Problems.” In 2016 IEEE International Conference on Big Journal of Advanced Manufacturing Technology 91 (5):
Data (Big Data), Washington D.C., USA, 2046–2050. IEEE. 1879–1887.
Peres, R., A. D. Rocha, J. P. Matos, and J. Barata (2018). Go0dman Tambe, P. P., and M. S. Kulkarni. 2015. “A Superimposition
Data Model-Interoperability in Multistage Zero Defect Based Approach for Maintenance and Quality Plan
Manufacturing. In 2018 IEEE 16th International Conference on Optimization with Production Schedule, Availability, Repair
Industrial Informatics (INDIN), Porto, Portugal, 815–821. IEEE. Time and Detection Time Constraints for a Single Machine.”
Prat, N. 2019, Jun. ”Augmented Analytics.” Business & Journal of Manufacturing Systems 37: 17–32.
Information Systems Engineering 61 (3): 375–380. Teti, R. 2015. “Advanced IT Methods of Signal Processing and
Psarommatis Giannakopoulos, F., S. Prouvost, G. May, and Decision Making for Zero Defect Manufacturing in
D. Kyritsis. 2020b. “Product Quality Improvement Policies Machining.” Procedia Cirp 28: 3–15.
in Industry 4.0: Characteristics, Enabling Factors, Barriers, Teucke, M., E. Broda, A. Boerold, and M. Freitag. 2018. “Using
and Evolution Toward Zero Defect Manufacturing.” Sensor-Based Quality Data in Automotive Supply Chains.”
Frontiers in Computer Science 2: 26. Machines 6 (4): 53.
Psarommatis, F., and D. Kiritsis. 2018. “A Scheduling Tool for Tidriri, K., N. Chatti, S. Verron, and T. Tiplica. 2016. “Bridging
Achieving Zero Defect Manufacturing (ZDM): A Conceptual Data-Driven and Model-Based Approaches for Process Fault
Framework.” In IFIP International Conference on Advances in Diagnosis and Health Monitoring: A Review of Researches
Production Management Systems, 271–278. Cham: Springer. and Future Challenges.” Annual Reviews in Control 42: 63–81.
Psarommatis, F., G. May, P. A. Dreyfus, and D. Kiritsis. 2020a. Tranfield, D., D. Denyer, and P. Smart. 2003. “Towards
“Zero Defect Manufacturing: State-Of-The-Art Review, a Methodology for Developing Evidence-informed
Shortcomings and Future Directions in Research.” Management Knowledge by Means of Systematic Review.”
International Journal of Production Research 58 (1): 1–17. British Journal of Management 14 (3): 207–222.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 701

Tsai, C. W., C. F. Lai, H. C. Chao, and A. V. Vasilakos. 2015. Wilhelm, Y., P. Reimann, W. Gauchel, and B. Mitschang. 2021.
“Big Data Analytics: A Survey.” Journal of Big Data 2 (1): “Overview on Hybrid Approaches to Fault Detection and
1–32. Diagnosis: Combining Data-Driven, Physics-Based and
Vater, J., L. Harscheidt, and A. Knoll. 2019, March. Smart Knowledge-Based Models.” Procedia Cirp 99: 278–283.
Manufacturing with Prescriptive Analytics. In 2019 8th Wuest, T., C. Irgens, and K. D. Thoben. 2014. “An Approach to
International Conference on Industrial Technology and Monitoring Quality in Manufacturing Using Supervised
Management (ICITM), Cambridge, United Kingdom, Machine Learning on Product State Data.” Journal of
224–228. IEEE. Intelligent Manufacturing 25 (5): 1167–1180.
Wang, K. S. 2013. “Towards Zero-Defect Manufacturing (Zdm) Wuest, T., D. Weimer, C. Irgens, and K. D. Thoben. 2016.
—a Data Mining Approach.” Advances in Manufacturing “Machine Learning in Manufacturing: Advantages,
1 (1): 62–74. Challenges, and Applications.” Production & Manufacturing
Wang, G., A. Ledwoch, R. M. Hasani, R. Grosu, and A. Brintrup. Research 4 (1): 23–45.
2019. “A Generative Neural Network Model for the Quality Zonnenshain, A., and R. S. Kenett. 2020. “Quality 4.0—the
Prediction of Work in Progress Products.” Applied Soft Challenging Future of Quality Engineering.” Quality
Computing 85: 105683. Engineering 32 (4): 614–626.
Wei, Y., D. Wu, and J. Terpenny. 2020. “Decision-Level Data Zonta, T., C. A. da Costa, R. da Rosa Righi, M. J. de Lima, E. S. da
Fusion in Quality Control and Predictive Maintenance.” IEEE Trindade, and G. P. Li. 2020. Predictive Maintenance in the
Transactions on Automation Science and Engineering 18 (1): Industry 4.0: A Systematic Literature Review. Computers &
184–194. Industrial Engineering, Vol. 150.