Hiba El Ouajdi*
| Omar Fanidi | Mustapha Hain
© 2026 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
OPEN ACCESS
The adoption of Predictive Maintenance (PdM) using Artificial Intelligence (AI) and Internet of Things (IoT) has been accelerated by Industry 4.0, however, it is still challenging to implement it for Small and Medium-sized Enterprises (SMEs) in data-scarce environments. This systematic review investigates the shift from reactive to proactive maintenance models, taking into account the particular limitations of small datasets. In contrast to current surveys that focus on big data analytics, this study evaluates the effectiveness of machine learning (ML) algorithms under data sparsity. We find that deep learning (DL) is the ideal approach for large datasets but classical algorithms such as Support Vector Machines (SVM) and Random Forests often provide better reliability for SMEs where historical failure data is limited. Moreover, we highlight low-cost sensor integration and federated learning as key enablers for SME digital transformation. The review ends with a roadmap for cross-machine model generalization and overcoming the socio-technical challenges to PdM adoption.
Proactive Maintenance, Industry 4.0, Artificial Intelligence, data scarcity, Small and Medium-sized Enterprises, deep learning
The emergence of Industry 4.0 has introduced a new era of technological integration and industrial growth, setting the stage for more advanced manufacturing trends [1, 2]. Within this context, the industrial landscape is undergoing a fundamental shift from traditional reactive and preventive maintenance to intelligent Predictive Maintenance (PdM) [3]. Traditional strategies, despite their simplicity, can often lead to unplanned downtime or unnecessary maintenance costs due to rigid scheduling [4]. The emergence of Industry 4.0, which is defined by the integration of cyber physical systems (CPS), the Internet of Things (IoT), and Artificial Intelligence (AI), has provided the technical foundation for real-time equipment monitoring [5, 6]. PdM utilizes continuous streams of sensor data to predict equipment failures ahead of time, thus optimizing the Remaining Useful Life (RUL) of assets [7-9]. However, the “big data” assumption, which is dominant in the current literature, is not representative of the reality of many industrial sites, especially Small and Medium-sized Enterprises (SMEs) [10].
Such environments often have data scarcity due to legacy equipment, high costs of deploying sensors, and the rare occurrence of critical failures [11]. We present a systematic analysis of PdM strategies for data-deficient environments in this paper. We bridge the world of high-level Industry 4.0 concepts and the concrete implementation issues. In particular, we investigate the performance of AI algorithms trained on limited data, and we propose a socio-technical framework to allow SMEs to implement PdM without prohibitive upfront investments [12, 13].
Unanticipated equipment failures can lead to significant financial loss, safety risks and environmental damages, especially for SMEs with limited resources [14]. Industry 4.0, despite its potential to improve efficiency and productivity, often lacks a practical connection to the operational realities of smaller companies that face outdated infrastructure, limited capital for new technologies, and a lack of specialized technical expertise [15]. This review is not confined to fault detection, but also discusses the wider implications of PdM for operational optimization, resource allocation, and long-term sustainability [16]. This systematic review contributes uniquely to the literature by focusing on data-scarce environments and the specific needs of SMEs, differentiating itself from the existing surveys that mostly focus on data-rich industrial contexts [17]. Our work aims to provide actionable insights and a clear roadmap for SMEs to successfully implement AI-driven PdM solutions, fostering resilience and competitive advantage in the ever-changing industrial landscape.
This review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure transparency and reproducibility [18, 19].
2.1 Search strategy
Our study involved conducting a thorough search through various scientific database sources such as IEEE Xplore, ScienceDirect, Scopus, Web of Science, and ACM Digital Library. These database sources were chosen because of the huge volume of research papers that covered engineering, computer science, and industrial management literature with regard to Industry 4.0, AI, and PdM studies [18, 19]. The timeframe of the search is limited from January 2014 to December 2024 in order to ensure the most recent progress made in Industry 4.0 and AI-based PdM, consistent with the rapid changes in these domains, since the appearance of Industry 4.0 definitions [20]. The search query involved the use of Boolean operators to link the key terms associated with maintenance, technology, and data limitations:
(Predictive Maintenance OR PdM) AND (Industry 4.0 OR IIoT OR Smart Manufacturing) AND (AI OR Machine Learning (ML) OR Deep Learning (DL)) AND (Data Scarcity OR Limited Data OR Small Data OR SMEs).
2.2 PRISMA flowchart and selection process
The selection process was based on four steps: identification, screening, eligibility, and inclusion.
1. Identification: An initial search resulted in 425 references (Figure 1 and Table 1).
2. Screening: After eliminating duplicates and screening titles and abstracts, 112 articles were left.
3. Eligibility: The full-text review helped to exclude 66 articles that either did not deal with the topic of data-scarce environments or lacked empirical validation of AI.
4. Inclusion: In total, there were 46 articles used for the final synthesis, in addition to 14 important studies from 2020 to 2025 [21].
Figure 1. PRISMA flow diagram of the systematic literature review (SLR) process
Table 1. Core literature analysis
|
Ref. |
Core AI Algorithms |
Key Findings / Contribution |
|
[22] |
SLR / Multiple |
Comprehensive review of PdM in Industry 4.0. |
|
[23] |
CPS / AI-based Systems |
Human-centered design in Industry 4.0 environments. |
|
[24] |
AI + Blockchain |
Using pre-trained Blockchain integration in Industry 4.0 systems. |
|
[25] |
ML / Data Analysis |
Economic and social impact of Industry 4.0. |
|
[26] |
AI / Trustworthy AI |
AI impact on sustainability in Industry 4.0. |
|
[27] |
AI-based Decision Support |
Low-power Strategic roadmap for Industry 4.0 transformation. |
Industry 4.0 does not refer to simply having tools, but rather to a transformation towards an age where connectivity leads to intelligence and seeks to form smart factories through integrating CPS, IoT, and AI [28]. Nevertheless, the evolution from one state to another faces challenges, particularly the “data rich but information poor” paradox [29]. The paradox indicates the presence of large volumes of data in industrial settings, which need to be converted to insights to inform decision-making.
Without the proper analytical skills, however, the amount of data generated, in addition to its heterogeneity, velocity, and veracity, makes it difficult to exploit for complex applications such as PdM [30]. In general, there are six basic principles that govern the concept of Industry 4.0 and allow for the creation of smart factories [31]:
Interoperability: Ability of CPS, people, and smart factories to network [32]. Essential for PdM to ensure data flow among PdM components.
Virtualization: The development of virtual models of the physical environment, such as machines, products, and manufacturing processes, using sensors and digital twins [33]. It provides for monitoring and predicting possible failure scenarios.
Decentralization: The ability of CPS to act independently, minimizing the dependence on centralized control and increasing the speed of reaction to possible failures [34].
Real-Time Feature: Collecting and analyzing data immediately, making it possible to recognize problems and take action in PdM [35].
Service-Oriented Approach: Providing services (internal and external) via cloud computing technologies to develop a flexible and scalable solution for PdM, especially for SMEs [36].
Modularity: The ability to add and remove separate modules when required to create customizable PdM solutions for various manufacturing settings [37].
However, despite all the basic tenets of Industry 4.0, its implementation within already established infrastructures poses many obstacles. Many factories run on outdated machinery that does not support any kind of digital connection, hence making the process of obtaining and integrating data quite problematic [38]. Lack of standardization, cybersecurity problems, and high capital outlays further aggravate the situation [39] (Table 2).
Table 2. Key enabling technologies in Industry 4.0 for PdM
|
Technology |
Role in PdM |
Strategic Benefit |
Ref. |
|
Artificial Intelligence (AI) |
Core predictive engine (ML, DL, Reinforcement). |
Improves RUL prediction accuracy and maintenance decision-making. |
[40] |
|
Cyber-Physical Systems (CPS) |
Integration of physical assets with digital twins. |
Enables real-time monitoring for ML/DL-based PdM systems. |
[41] |
|
Internet of Things (IoT) |
Network of sensors for real-time data acquisition. |
Provides high-quality data for RUL estimation and failure prediction. |
[42] |
|
Big Data Analytics |
Processing large-scale heterogeneous datasets. |
Supports ML/DL models for detecting complex failure patterns. |
[43] |
|
Cloud & Edge Computing |
Infrastructure for data storage and local processing. |
Enables scalable and low-latency deployment of PdM models |
[44] |
|
Digital Twin |
High-fidelity virtual replica of industrial assets. |
Enhances RUL prediction through simulation and real-time updates. |
[45] |
Furthermore, the emphasis laid by the concept of Industry 4.0 on Big Data is not relevant to smaller companies and data-deficient contexts, where the amount of data available is always limited [40].
An important gap noticed in existing literature [46] is the absence of an appropriate bridge between generic technologies of Industry 4.0 and concrete demands of PdM technology. Here, IoT and Cloud Computing form the sensory and nervous systems, while AI forms the brain system. For SMEs, the issue with such a bridge is that it remains difficult due to exorbitant costs involved in the use of gold standard sensors. Thus, virtualization and modularity should be focused upon [47].
The choice of AI algorithm is the most vital step in developing a PdM system. Although currently, DL is preferred, it relies heavily on vast amounts of labeled data, which are not always available in most industrial settings [48].
5.1 Comparison between classical ML and DL methods
“Data-scarce” cases generally favor classical ML models over deep neural networks. Specifically, algorithms like SVM and RF have shown remarkable performance on small datasets owing to their resistance to overfitting and reduced computational cost [49]. In this regard, SVM is especially effective in high-dimensional spaces with a few data points because it can find an optimal hyperplane separating faulty states from healthy ones using only a few support vectors [49]. On the other hand, the use of ensemble methods in RF has enabled the algorithm to reduce variance and be robust in its predictions, making RF ideal for small datasets that contain either limited features or instances [49]. Gradient Boosting Machine (GBM) and Extreme Gradient Boosting (XGBoost) are other examples of such algorithms with outstanding performance in “data-scarce” cases.
On the contrary, DL architectures, like Convolutional Neural Networks (CNNs) or Long Short-Term Memory (LSTMs), need thousands of failure samples to capture the degradation patterns and generalize well [50]. Being complicated in nature with lots of parameters, DL models are extremely prone to overfitting, especially when the available training dataset is very small. For many SMEs, a machine may experience just one or two failures in a year, hence it would be impractical to build a DL model from scratch without performing massive amounts of data augmentation or generating new data, which poses another challenge [51]. Moreover, the computation power required for training such algorithms is not always feasible for SMEs that do not have adequate IT resources. To solve this problem, Transfer Learning techniques have been developed. With Transfer Learning, one can adapt the model trained in a rich data environment (e.g. a similar machine from another factory or public dataset) by applying the data available in the target SMEs through fine-tuning [51]. Examples of such applications can be found in the work [51]. In addition to the aforementioned methods, other approaches to PdM such as the One-Class SVM and Isolation Forest can be effectively employed in conditions where there is a lack of data, since they rely on learning about healthy cases and detecting any anomalies as possible defects, bypassing the necessity of having substantial data on failures.
5.2 Performance metrics for low-data environments
In addition to accuracy, performance of PdM solutions in SMEs also needs to be assessed in terms of their reliability and intelligibility [52]. An operator would trust an easy-to-understand decision tree that provides reasoning behind a decision to perform maintenance better than a “black box” neural network. According to recent studies, hybrid methods using both physics and ML are most effective in data-scarce environments [53].
5.3 Effective strategies and solutions for Small and Medium-Sized Enterprises
The abstract mentions "effective strategies for transition"; however most approaches are still conceptual. In the case of SMEs, the key obstacles are high costs of initial investment, lack of expertise in AI technologies, and the need for replacement of their existing infrastructure [54].
5.4 Low-cost solution framework
We suggest that the three-level strategy is suitable for overcoming these problems, and each step has its own advantages, although there are some obstacles to be overcome as well:
1. Low-Cost Sensors Implementation: Instead of making expensive upgrades and purchasing new equipment, SMEs may use retro-fitting solutions based on the "plug-and-play" sensors. Technologies like Micro-Electro-Mechanical System (MEMS) accelerometers, acoustic sensors, thermal cameras in conjunction with Low-Power Wide-Area Network (LPWAN), such as Long Range Wide Area Network (LoRaWAN) and Narrowband Internet of Things (NB-IoT) allow to send data without the necessity of creating complex infrastructure [55]. Moreover, using such sensors may result in reducing hardware costs up to 70% when compared to standard industrial devices. Nevertheless, the issue of ensuring data quality becomes crucial.
2. Cloud-based SaaS Models (PdMaaS): The cloud-based SaaS solution, referred to as PdM as a Service (PdMaaS), gives SMEs access to advanced AI algorithms and powerful cloud infrastructure that hosts the system. Using the cloud-based model, the SMEs do not require any huge upfront investment towards hardware and AI professionals since SMEs are paying for the use of this service via monthly subscriptions [56]. Some notable strengths of the PdMaaS include the ability to scale, less burden on operations, and availability of state-of-the-art AI systems. However, the risk of losing data privacy, vendor lock-in, and unreliable internet connection must be considered.
3. Federated Learning: Federated learning is an advanced distributed learning method where multiple SMEs can collectively build one model based on the data from all the SMEs. In federated learning, the data from SMEs are never shared, but the model parameters (weight parameters) from each local model are combined together for better performance of the global model. This is known as knowledge pooling, which helps to resolve the issue of individual lack of sufficient data. However, federated learning poses challenges in terms of aggregating models and managing communication overhead and fairness.
5.5 Socio-technological integration
PdM implementation requires more than just mastering technology; it is a major organizational issue. Small organizations should develop a corporate culture of decision-making based on data analysis, which implies a radical change of mentality and operation processes [57]. In particular, it is important to ensure the existence of training programs that will provide maintenance personnel with relevant digital skills for cooperation with AI. The main task is not to create a fully automated system, but to enable people with the ability to independently verify the results of the work of AI, using their extensive experience and intuitive awareness of the machinery [58]. Management practices play an important role in overcoming any resistance to changes associated with new technologies and smooth integration of PdM with them into current processes [59].
As more advanced AI-based PdM models are developed, a number of issues will need to be addressed to ensure that their use becomes widely accepted within the industry as a whole.
Dealing with Concept Drift: Since industrial processes change continuously (for instance, changes in material composition, temperature, and machinery wear), the development of effective AI models must account for concept drift, changes in environmental conditions that do not necessitate complete retraining [59].
Cross-Machine PdM: Another issue that prevents the broad application of PdM is the inability of current models to generalize across machines (regardless of brand). Therefore, one potential direction of future research could be finding ways to train universal failure predictors that would apply to multiple brands and models of machinery [60].
Sustainability & Green PdM: Besides its economic value, PdM plays an essential part in ensuring industrial sustainability. It is important for researchers to find out how the use of PdM can lead to less CO2 emissions and a reduction in energy waste, making the manufacturing process more sustainable by utilizing resources better and increasing the life cycle of machines. This also involves the study of the environmental effects of using AI-based models.
In any case, this systematic review has shown an evident discrepancy between the idea of Industry 4.0 and real-life problems experienced by a majority of SMEs. While a "Big Data" approach is considered a prerequisite for intelligent manufacturing and PdM, it should be borne in mind that many SMEs are forced to work in circumstances where there is a lack of data, financial limitations, and a deficiency of expertise in digitalization. Therefore, solutions tailored to large-scale enterprises with large amounts of data and digital technologies cannot always be applied in a smaller industrial context.
According to the results of this review, it is clear that PdM can function perfectly without the need to collect huge amounts of data. In this regard, classical ML models like SVM are still quite relevant because of their effectiveness, high interpretability, and performance even when working with small amounts of data. Moreover, new trends in ML such as Transfer and Federated Learning provide great potential in terms of solving a data-scarce problem while maintaining privacy and data ownership.
Furthermore, the review highlights the significance of cost-efficient technological infrastructures. Affordable sensors, edge computing and low-cost Software-as-a-Service (SaaS) platforms dramatically lower costs associated with deploying PdM practices by making it possible to benefit from advanced analysis and AI-assisted maintenance techniques without the need for considerable spending on hardware equipment, computing power, and qualified specialists. Thus, the democratization of digital innovations is necessary to bring the advantages of Industry 4.0 to the benefit of small and medium-sized enterprises.
Moreover, besides considering technology, this research highlights the fact that the implementation of PdM requires a holistic transformation which encompasses both technical and social aspects. AI should serve as an instrument that helps to make better decisions but not as something that replaces human experience altogether. Humans are needed to evaluate and interpret the results obtained during model execution, validate the maintenance decisions made according to the data collected and manage any abnormal situations.
The adoption of intelligent maintenance systems is the final aspect in which SMEs can benefit from opportunities that lie ahead in adopting PdM. The implementation of such systems will lead to greater reliability of equipment, reduced downtime, improved optimization of maintenance costs, and greater environmental sustainability due to improved resource management. Research in the future should be focused on the development of lighter AI models, standardized approaches to data sharing, and explanations of AI systems, all of which can help SMEs in their needs.
It is a pleasure for the authors to extend their sincere thanks to the Laboratory of Intelligent, Industrial and Mechanical Systems Engineering (LISIME), ENSAM Casablanca, Hassan II University of Casablanca, Morocco, for offering the intellectual atmosphere that helped in accomplishing the present research study.
|
PdM |
Predictive Maintenance |
|
AI |
Artificial Intelligence |
|
IoT |
Internet of Things |
|
SMEs |
Small and Medium-sized Enterprises |
|
CPS |
Cyber-Physical Systems |
|
RUL |
Remaining Useful Life |
|
ML |
Machine Learning |
|
DL |
Deep Learning |
|
SVM |
Support Vector Machines |
|
RF |
Random Forests |
|
PRISMA |
Preferred Reporting Items for Systematic Reviews and Meta-Analyses |
|
CNNs |
Convolutional Neural Networks |
|
LSTMs |
Long Short-Term Memory Networks |
|
GBM |
Gradient Boosting Machine |
|
XGBoost |
Extreme Gradient Boosting |
|
MEMS |
Micro-Electro-Mechanical Systems |
|
LPWAN |
Low-Power Wide-Area Network |
|
LoRaWAN |
Long Range Wide Area Network |
|
NB-IoT |
Narrowband Internet of Things |
|
PdMaaS |
Predictive Maintenance as a Service |
[1] Xu, L.D., Xu, E.L., Li, L. (2018). Industry 4.0: State of the art and future trends. International Journal of Production Research, 56(8): 2941-2962. https://doi.org/10.1080/00207543.2018.1444806
[2] Rüßmann, M., Lorenz, M., Gerbert, P., Waldner, M., Justus, J., Engel, P., Harnisch, M. (2015). Industry 4.0: The future of productivity and growth in manufacturing industries. Boston Consulting Group, 9(1): 54-89. https://www.marioagostinelli.it/wp-content/uploads/2017/07/09_Industry-4.0-The-Future-of-Productivity-and-Growth-in-Manufacturing-Industries.pdf.
[3] Wan, J.F., Cai, H., Zhou, K.L. (2015). Industrie 4.0: Enabling technologies. In Proceedings of 2015 International Conference on Intelligent Computing and Internet of Things, Harbin, China, pp. 135-140, https://doi.org/10.1109/ICAIOT.2015.7111555
[4] Lasi, H., Fettke, P., Kemper, H.G., Feld, T., Hoffmann, M. (2014). Industry 4.0. Business & Information Systems Engineering, 6(4): 239-242.https://doi.org/10.1007/s12599-014-0334-4
[5] Hermann, M., Pentek, T., Otto, B. (2016). Design principles for Industrie4.0scenarios. In 2016 49th Hawaii International Conference on System Sciences (HICSS), Koloa, HI, USA, pp. 3928-3937. https://doi.org/10.1109/HICSS.2016.488
[6] Oztemel, E., Gursev, S. (2018). Literature review of Industry 4.0 and related technologies. Journal of Intelligent Manufacturing, 31(1): 127-182. https://doi.org/10.1007/s10845-018-1433-8
[7] Oks, S.J., Jalowski, M., Lechner, M., Mirschberger, S., Merklein, M., Vogel-Heuser, B., Möslein, K.M. (2024). Cyber-physical systems in the context of Industry 4.0: A review, categorization and outlook. Information Systems Frontiers, 26(5): 1731-1772. https://doi.org/10.1007/s10796-022-10252-x
[8] Szymanska, O., Adamczak, M., Cyplik, P. (2018). Logistics4.0-a new paradigm or set of known solutions? Research in Logistics & Production, 7(4): 299-310.
[9] Carvalho, T.P., Soares, F.A.A.M.N., Vita, R., Francisco, R.P., Basto, J.P., Alcalá, S.G.S. (2019). A systematic literature review of machine learning methods applied to predictive maintenance. Computers & Industrial Engineering, 137: 106024. https://doi.org/10.1016/j.cie.2019.106024
[10] Grubisic, V.V.F., Aguiar, J.P.F., Simeu-Abazi, Z. (2020). A review on intelligent predictive maintenance: Bibliometric analysis and new research directions. In 2020 International Conference on Control, Automation and Diagnosis (ICCAD), Paris, France, pp. 1-6, https://doi.org/10.1109/ICCAD49821.2020.9260504
[11] Peres, R.S., Jia, X.D., Lee, J., Sun, K.Y., Colombo, A.W., Barata, J. (2020). Industrial artificial intelligence in Industry 4.0-systematic review, challenges and outlook. IEEE Access, 8: 220121-220139. https://doi.org/10.1109/ACCESS.2020.3042874
[12] Bertolini, M., Mezzogori, D., Neroni, M., Zammori, F. (2021). Machine learning for industrial applications: A comprehensive literature review. Expert Systems with Applications, 175: 114820. https://doi.org/10.1016/j.eswa.2021.114820
[13] Noman, M.A., Nasr, E.S.A., Al-Shayea, A., Kaid, H. (2019). Overview of predictive condition based maintenance research using bibliometric indicators. Journal of King Saud University-Engineering Sciences, 31(4): 355-367. https://doi.org/10.1016/j.jksues.2018.02.003
[14] Sajid, S., Haleem, A., Bahl, S., Javaid, M., Goyal, T., Mittal, M. (2021). Data science applications for predictive maintenance and materials science in context to Industry 4.0. Materials Today: Proceedings, 45: 4898-4905. https://doi.org/10.1016/j.matpr.2021.01.357
[15] Sharma, A., Pandey, H. (2019). Big data and analytics in Industry 4.0. In a Roadmap to Industry 4.0: Smart Production, Sharp Business and Sustainable Development, pp. 57-72. https://doi.org/10.1007/978-3-030-14544-6_4
[16] Hossain, M.S., Muhammad, G. (2016). Cloud-assisted Industrial Internet of Things (IIoT)–Enabled framework for health monitoring. Computer Networks, 101: 192-202. https://doi.org/10.1016/j.comnet.2016.01.009
[17] Abayadeera, M.R., Ganegoda, G.U. (2024). Digital twin technology: A comprehensive review. International Journal of Scientific Research and Engineering Trends, 10(4): 1485-1504.
[18] Jones, D., Snider, C., Nassehi, A., Yon, J., Hicks, B. (2020). Characterising the Digital Twin: A systematic literature review. CIRP Journal of Manufacturing Science and Technology, 29: 36-52. https://doi.org/10.1016/j.cirpj.2020.02.002
[19] Moeuf, A., Pellerin, R., Lamouri, S., Tamayo-Giraldo, S., Barbaray, R. (2018). The industrial management of SMEs in the era of Industry 4.0. International Journal of Production Research, 56(3): 1118-1136. https://doi.org/10.1080/00207543.2017.1372647
[20] Horváth, D., Szabó, R.Z. (2019). Driving forces and barriers of Industry 4.0: Do multinational and small and medium-sized companies have equal opportunities? Periodica Polytechnica Social and Management Sciences, 27(2): 146-158. https://doi.org/10.3311/PPso.12664
[21] Kamble, S.S., Gunasekaran, A., Gawankar, S.A. (2018). Industry 4.0 and the digital supply chain: A systematic review. International Journal of Production Economics, 204: 408-425. https://doi.org/10.1016/j.ijpe.2018.09.018.
[22] Sony, M., Naik, S. (2019). Critical factors for the successful implementation of Industry 4.0: A review and future research direction. Production Planning & Control, 31(10): 799-815. https://doi.org/10.1080/09537287.2019.1691278
[23] Fantini, P., Pinzone, M., Taisch, M. (2018). Placing the operator at the centre of Industry 4.0 design: Modelling and assessing human activities within cyber-physical systems. Computers & Industrial Engineering, 125: 403-415. https://doi.org/10.1016/j.cie.2018.01.025
[24] da Rosa Righi, R., Alberti, A.M., Singh, M. (2020). Blockchain Technology for Industry 4.0. Springer Singapore.
[25] Cezanne, C., Lorenz, E., Saglietto, L. (2020). Exploring the economic and social impacts of Industry 4.0. Revue D'économie Industrielle, 169: 11-35. https://doi.org/10.4000/rei.8643
[26] Muhammad, D.S., Muhammad, W., Anis, F. (2025). Impact of Industry 4.0 on sustainability. Applied Mechanics and Materials, 929: 175-184. https://doi.org/10.4028/p-EQ3fhV
[27] Ghobakhloo, M. (2018). The future of manufacturing industry: A strategic roadmap toward Industry 4.0. Journal of Manufacturing Technology Management, 29(6): 910-936. https://doi.org/10.1108/JMTM-02-2018-0057
[28] Sony, M., Naik, S. (2020). Critical review of lean manufacturing and Industry 4.0 integration. Journal of Manufacturing Technology Management, 31(4): 749-774. https://doi.org/10.1108/JMTM-01-2019-0031.
[29] Schumacher, A., Erol, S., Sihn, W. (2016). A maturity model for assessing Industry 4.0 readiness and maturity of manufacturing enterprises. Procedia CiRP, 52: 161-166. https://doi.org/10.1016/j.procir.2016.07.040
[30] Orellana, F., Torres, R. (2019). From legacy-based factories to smart factories level 2 according to the Industry 4.0. International Journal of Computer Integrated Manufacturing, 32(4-5): 441-451. https://doi.org/10.1080/0951192X.2019.1609702
[31] Brettel, M., Klein, M., Friederichsen, N. (2014). The relevance of manufacturing flexibility in the context of Industrie 4.0. Procedia CiRP, 41: 105-110. https://doi.org/10.1016/j.procir.2015.12.047
[32] Fischer, J., Obst, B., Lee, B. (2017). Integrating material flow simulation tools in a service-oriented industrial context. In 2017 IEEE 15th International Conference on Industrial Informatics (INDIN), Emden, Germany, pp. 1135-1140, https://doi.org/10.1109/INDIN.2017.8104933
[33] Anderl, R. (2015). Industrie 4.0–Technological approaches, use cases, and implementation. at-Automatisierungstechnik, 63(10): 753-765. https://doi.org/10.1515/auto-2015-0025
[34] Adedoyin, F.F., Bekun, F.V., Driha, O.M., Balsalobre-Lorente, D. (2020). The effects of air transportation, energy, ICT and FDI on economic growth in the Industry 4.0 era: Evidence from the United States. Technological Forecasting and Social Change, 160: 120297. https://doi.org/10.1016/j.techfore.2020.120297
[35] Porter, M. E., Heppelmann, J.E. (2014). How smart, connected products are transforming competition. Harvard Business Review, 92(11): 64-88. https://moodle2.units.it/pluginfile.php/727297/mod_resource/content/1/How-smart-connected-products-are-transforming-competition.pdf.
[36] Trotta, D., Garengo, P. (2018). Industry 4.0 key research topics: A bibliometric review. In 2018 7th International Conference on Industrial Technology and Management (ICITM), Oxford, UK, pp. 113-117, https://doi.org/10.1109/ICITM.2018.8333930
[37] Souza, M.L.H., da Costa, C.A., de Oliveira Ramos, G., da Rosa Righi, R. (2020). A survey on decision-making based on system reliability in the context of Industry 4.0. Journal of Manufacturing Systems, 56: 133-156. https://doi.org/10.1016/j.jmsy.2020.05.016
[38] Madenoğlu, F.S. (2023). Industry 4.0: A bibliometric analysis.
[39] Mariani, M., Borghi, M. (2019). Industry 4.0: A bibliometric review of its managerial intellectual structure and potential evolution in the service industries. Technological Forecasting and Social Change, 149: 119752. https://doi.org/10.1016/j.techfore.2019.119752
[40] Salkin, C., Oner, M., Ustundag, A., Cevikcan, E. (2017). A conceptual framework for Industry 4.0. In Industry 4.0: Managing the Digital Transformation, pp. 3-23. https://doi.org/10.1007/978-3-319-57870-5_1
[41] Al Zadjali, S., Ullah, A. (2021). Impact of Industry 4.0 in manufacturing sector. International Journal of Management Science and Business Administration, 7(6): 25-33. https://doi.org/10.18775/ijmsba.1849-5664-5419.2014.76.1003
[42] Neumann, W.P., Winkelhaus, S., Grosse, E.H., Glock, C.H. (2021). Industry 4.0 and the human factor: A systems framework and analysis methodology for successful development. International Journal of Production Economics, 233: 107992. https://doi.org/10.1016/j.ijpe.2020.107992
[43] Agostini, L., Nosella, A. (2020). The adoption of Industry 4.0 technologies in SMEs: Results of an international study. Management Decision, 58(4): 625-643. https://doi.org/10.1108/MD-09-2018-0973
[44] Haleem, A., Javaid, M. (2019). Additive manufacturing applications in Industry 4.0: A review. Journal of Industrial Integration and Management, 4(04): 1930001. https://doi.org/10.1142/S242486221930001
[45] Fatorachian, H., Kazemi, H. (2021). Impact of Industry 4.0 on supply chain performance. International Journal of Production Research, 59(23): 7047-7068. https://doi.org/10.1080/00207543.2020.1712487
[46] Mazzei, D., Ramjattan, R. (2022). Machine learning for Industry 4.0: A systematic review using deep learning-based topic modelling. Sensors, 22(22): 8641. https://doi.org/10.3390/s22228641
[47] Uçar, A., Karakose, M., Kırımça, N. (2024). Artificial intelligence for predictive maintenance applications: Key components, trustworthiness, and future trends. Applied Ciences, 14(2): 898. https://doi.org/10.3390/app14020898
[48] Cheng, X., Chaw, J.K., Goh, K.M., Ting, T.T., Sahrani, S., Ahmad, M.N., Kadir, R.A., Ang, M.C. (2022). Systematic literature review on visual analytics of predictive maintenance in the manufacturing industry. Sensors, 22(17): 6321. https://doi.org/10.3390/s22176321
[49] Azari, M.S., Flammini, F., Santini, S., Caporuscio, M. (2023). A systematic literature review on transfer learning for predictive maintenance in Industry 4.0. IEEE Access, 11: 12887-12910. https://doi.org/10.1109/ACCESS.2023.3239784
[50] Zonta, T., da Costa, C.A., da Rosa Righi, R., de Lima, M.J., da Trindade, E.S., Li, G.P. (2020). Predictive maintenance in the Industry 4.0: A systematic literature review. Computers & Industrial Engineering, 150: 106889. https://doi.org/10.1016/j.cie.2020.106889
[51] Krupitzer, C., Wagenhals, T., Züfle, M., Lesch, V., Schäfer, D., Mozaffarin, A., Edinger, J., Becker, C., Kounev, S. (2020). A survey on predictive maintenance for Industry 4.0. arXiv preprint, arXiv:2002.08224. https://doi.org/10.48550/arXiv.2002.08224
[52] Zhu, T.W., Ran, Y.Y., Zhou, X., Wen, Y.G. (2019). A survey of predictive maintenance: Systems, purposes and approaches. arXiv preprint, arXiv:1912.07383. https://doi.org/10.1109/COMST.2025.3567802
[53] Liao, Y., Deschamps, F., Loures, E.F.R., Ramos, L.F P. (2017). Past, present and future of Industry 4.0 – A systematic literature review and research agenda proposal. International Journal of Production Research, 55(12): 3609-3629. https://doi.org/10.1080/00207543.2017.1308576
[54] Windmann, A., Wittenberg, P., Schieseck, M., Niggemann, O. (2024). Artificial intelligence in Industry 4.0: A review of integration challenges for industrial systems. In 2024 IEEE 22nd International Conference on Industrial Informatics (INDIN), Beijing, China, pp. 1-8. https://doi.org/10.1109/INDIN58382.2024.10774364
[55] Vasilache, A., Nitzsche, S., Floegel, D., Schuermann, T., von Dosky, S., Bierweiler, T., Mußler, M., Kälber, F., Hohmann, S., Becker, J. (2024). Low-power vibration-based predictive maintenance for Industry 4.0 using neural networks: A survey. In Machine Learning and Principles and Practice of Knowledge Discovery in Databases, pp. 149-161. https://doi.org/10.1007/978-3-032-25308-8_10
[56] Abbas, A. (2024). AI for predictive maintenance in industrial systems. International Journal of Advanced Engineering Technologies and Innovations, 1(1): 31-51. https://www.researchgate.net/publication/386142847_AI_for_Predictive_Maintenance_in_Industrial_Systems.
[57] Pejić Bach, M., Topalović, A., Krstić, Ž., Ivec, A. (2023). Predictive maintenance in Industry 4.0 for SMEs: A decision support system case study using open-source software. Designs, 7(4): 98. https://doi.org/10.3390/designs7040098
[58] Lee, J., Davari, H., Singh, J., Pandhare, V. (2018). Industrial artificial intelligence for Industry 4.0-based manufacturing systems. Manufacturing Letters, 18: 20-23. https://doi.org/10.1016/j.mfglet.2018.09.002
[59] Guidotti, D., Pandolfo, L., Pulina, L. (2025). A systematic literature review of supervised machine learning techniques for predictive maintenance in Industry4.0. IEEE Access. https://doi.org/10.1109/ACCESS.2025.3578686
[60] Kamble, S.S., Gunasekaran, A., Sharma, R. (2018). Analysis of the Driving and dependence power of Barriers to Adopt Industry 4.0 in Indian Manufacturing Industry. Computers in Industry, 101: 107-119. https://doi.org/10.1016/j.compind.2018.06.004