Enabling human-centric manufacturing: A systematic review of human-machine collaboration in Industry 5.0

1.1. Research aim

This study aims to systematically review and synthesize existing scholarly literature on human-machine collaboration in Industry 5.0, with a focus on identifying key enabling technologies, collaboration paradigms, and implementation challenges that shape the transition toward human-centric manufacturing systems.

1.2. Research objectives, formulated research questions, and identified gaps in literature

Table 1 presents a structured alignment between the study’s research objectives, corresponding research questions, and the key research gaps identified from the reviewed literature. By explicitly linking each aim of a specific research question, the table clarifies the study’s analytical focus and ensures methodological coherence. Additionally, the identified research gaps highlight critical limitations in existing Industry 5.0 research, thereby justifying the need for this review and guiding future investigations toward advancing human-centric manufacturing.

This study makes a distinct contribution to the field of human-machine collaboration in Industry 5.0 by addressing three interrelated objectives that collectively advance knowledge and practice. First, by identifying and categorizing key technological enablers, it provides a systematic framework that integrates AI, digital twins, cobots, and other technologies, moving beyond fragmented analyses to reveal how these tools collectively support human-centric manufacturing. Second, by analyzing dominant collaboration paradigms and their applications, the study links theoretical models, such as HMI, HRC, HMC, and HCPS, to practical manufacturing contexts, highlighting measurable outcomes related to trust, ergonomics, and worker well-being. Third, by synthesizing implementation challenges, including technological, organizational, and socio-ethical barriers, the research exposes critical gaps in empirical and longitudinal validation, especially in real industrial settings and developing economies. Together, these objectives offer a comprehensive, evidence-based roadmap for advancing human-centric Industry 5.0 adoption, bridging theory, practice, and policy considerations.

1.3. Literature review

2.1. Review design and PRISMA framework

This study employed a systematic literature review (SLR) explicitly guided by the PRISMA 2020 (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework to ensure methodological transparency, reproducibility, and completeness. PRISMA principles informed all stages of the review process, including literature identification, screening, eligibility assessment, inclusion, and synthesis.

The review was designed to address three predefined research objectives focused on identifying technological enablers, analyzing human-machine collaboration paradigms and applications, and synthesizing implementation challenges and research gaps in Industry 5.0 human-centric manufacturing. A structured review protocol was developed before data collection to minimize selection bias and ensure consistency in study identification and analysis.

2.2. Information sources and literature search strategy

Consistent with PRISMA recommendations for comprehensive coverage, a systematic search was conducted across multiple high-quality bibliographic databases to capture interdisciplinary research spanning manufacturing systems, robotics, AI, cyber–physical systems, and socio-technical studies relevant to Industry 5.0. The search strategy employed predefined keywords combined using Boolean operators, limited to peer-reviewed publications in English from 2017 to 2025, to reflect the emergence and maturation of Industry 5.0 research. In addition, reference lists of key review articles were screened to identify further relevant studies, in line with PRISMA guidance on supplementary search strategies. Table 2 summarizes the specific search strings applied across databases.

2.3. Eligibility criteria

Clear inclusion and exclusion criteria were established a priori in accordance with PRISMA. Studies were included if they explicitly addressed human-machine, human-robot, or human-cyber-physical collaboration, were situated within manufacturing or closely related industrial contexts, and examined enabling technologies, collaboration paradigms, applications, or implementation challenges aligned with Industry 5.0 human-centric principles. Conversely, studies were excluded if they focused solely on Industry 4.0 automation without human-centric considerations, addressed non-industrial domains exclusively, or lacked sufficient conceptual or methodological clarity relevant to the study objectives.

2.4. Study selection and screening process

The study selection process adhered to the PRISMA four-phase flow, comprising identification, screening, eligibility, and inclusion. In the identification phase, all records retrieved from the database searches were aggregated, and duplicate entries were removed to ensure accuracy. During the screening phase, titles and abstracts were examined to exclude studies that were clearly irrelevant. This initial screening was conducted by the first author, with cross-checks performed by co-authors to minimize bias.

In the eligibility phase, full-text articles of potentially relevant studies were assessed against the predefined inclusion and exclusion criteria. This step ensured that only studies directly addressing human–machine collaboration in Industry 5.0 manufacturing were considered. Finally, in the inclusion phase, sixty studies met the eligibility criteria and were incorporated into the final synthesis. These studies are comprehensively documented in Appendix A, providing traceability and transparency of inclusion decisions. The overall selection process is visually summarized in Fig. 1, which presents the PRISMA flow diagram.

Fig. 1PRISMA 2020

2.5. Quality appraisal of included studies

The credibility of the included evidence was evaluated in accordance with PRISMA recommendations through a qualitative quality appraisal of all selected studies. Unlike numerical scoring systems, this appraisal employed qualitative criteria, which are more appropriate for an interpretive systematic literature review. Each study was assessed based on its relevance to the research objectives, the clarity of its aims, the coherence of its methodology, and its contribution to theory or practice in human-centric manufacturing.

Given the exploratory and interdisciplinary nature of Industry 5.0 research, both empirical and conceptual studies were retained, provided they demonstrated analytical rigor and relevance. This inclusive approach ensured comprehensive coverage of the field while maintaining quality standards appropriate for a systematic review.

To enhance transparency, the quality appraisal was guided by four explicit criteria: (1) Relevance to Industry 5.0 human-machine collaboration, (2) Methodological rigor, including clarity of research design and data analysis, (3) Theoretical or practical contribution, particularly in advancing human-centric manufacturing, and (4) Clarity and coherence of reported findings.

Studies meeting these criteria were retained to ensure analytical robustness and alignment with the study objectives.

2.6. Data collection and extraction

A standardized data extraction form was developed before analysis, consistent with PRISMA. For each included study, the following data were systematically extracted: author(s) and year, study focus, enabling technologies discussed, type of human-machine collaboration, manufacturing application domain, key findings, and identified implementation challenges.

The extracted data formed the basis of Appendix A, which serves as the primary evidence table linking individual studies to the synthesized results presented in the main text.

2.7. Data organization, coding, and analysis

Extracted data were organized thematically and analyzed using a qualitative thematic synthesis approach. An inductive-deductive coding strategy was adopted, in which initial codes were derived from the research objectives, while additional themes emerged iteratively from the data. Studies were systematically coded and clustered into thematic categories covering technological enablers of human-machine collaboration, collaboration paradigms and manufacturing applications, as well as implementation challenges and research gaps. This structured coding process facilitated cross-study comparison and supported the development of integrative summary tables (Tables 2-4), in line with PRISMA guidance on transparent and reproducible data synthesis.

2.8. Synthesis of results

In accordance with PRISMA, a narrative synthesis was employed to integrate findings across heterogeneous study designs. This approach facilitated the identification of recurring patterns, convergent evidence, and unresolved issues in human-centric manufacturing for Industry 5.0.

The synthesis emphasized qualitative human-centric outcomes, such as trust, safety, ergonomics, cognitive workload, and well-being, alongside technological and operational considerations, reflecting the core principles of Industry 5.0.

2.9. Methodological rigor and transparency

Methodological rigor was reinforced through adherence to PRISMA guidelines, explicit documentation of search and screening procedures, and transparent reporting of inclusion decisions. The use of structured extraction templates, thematic coding, and cross-referencing between Appendix A and synthesized tables enhanced reliability and reproducibility.

Although this review did not involve quantitative meta-analysis due to heterogeneity in study designs and outcome measures, the systematic and transparent approach ensures robustness appropriate for a PRISMA-compliant qualitative SLR.

2.10. Tools and software used

Reference management was conducted using Mendeley, facilitating citation tracking and duplicate removal. Data extraction, coding, and synthesis were performed using Microsoft Excel and Microsoft Word to support structured qualitative analysis.

No bibliometric or automated text-mining tools were used, as the study prioritized interpretive depth and contextual understanding consistent with the human-centric orientation of Industry 5.0 research.

3.1. Identification and categorization of technological enablers of human-machine collaboration

The reviewed studies (Appendix A) reveal that human-centric manufacturing in Industry 5.0 is underpinned by a convergence of advanced digital, cognitive, and physical technologies rather than isolated automation tools. Analysis of the “technology/tools discussed” field across the dataset indicates five dominant technological clusters.

Table 3 summarizes key technological clusters supporting human-machine collaboration, detailing the enabling technologies, their roles in facilitating interaction, and the human-centric contributions they provide, including enhanced decision-making, safety, productivity, situational awareness, and adaptive, intuitive collaboration.

The analysis demonstrates that human-centric manufacturing emerges from the coordinated integration of intelligent cognition, digital representation, physical collaboration, and systemic orchestration rather than from isolated technological deployment. Artificial intelligence enables adaptive decision-making and contextual awareness; digital twins provide virtual representations of both human and machine states; collaborative robots support safe physical interaction; and human-cyber-physical systems orchestrate these elements in real time. Together, these technologies create synergistic, adaptive manufacturing environments that enhance productivity, safety, trust, and worker well-being while reinforcing the central role of humans in Industry 5.0 systems.

A critical observation emerging from the cluster analysis is the contrast between studies that treat AI capabilities and digital twin technologies as standalone performance enhancers versus those that examine their integrative potential. While studies such as [14] and [16] demonstrate discrete benefits of cobots and predictive maintenance, respectively, studies employing HCPS frameworks [13, 63] reveal distinctly superior human-centric outcomes when these clusters are combined, particularly in relation to adaptive task allocation and cognitive workload reduction. This divergence suggests that fragmented technology adoption, though common in practice, may systematically underperform compared to architecturally integrated deployments, a finding with direct implications for investment prioritization in SMEs.

3.2. Analysis of human-machine collaboration paradigms and manufacturing applications

Table 4 provides a comprehensive overview of human-machine collaboration paradigms in Industry 5.0 manufacturing, mapping each paradigm to its primary applications, associated human roles, and key human-centric outcomes. It highlights the progression from basic interface-level interactions to fully integrated, adaptive, and co-learning systems, demonstrating the expanding cognitive and creative involvement of humans. The framework emphasizes qualitative dimensions, such as trust, well-being, and skill development, alongside productivity, underscoring the shift toward human-centered, intelligent production ecosystems in modern manufacturing.

The matrix illustrates a clear evolution from basic interface-level interaction to advanced, adaptive, and co-learning collaboration models, reflecting Industry 5.0’s transition from purely efficiency-focused automation to human-centered value creation. Higher collaboration stages correspond to wider application areas, enriched cognitive roles for humans, and qualitative outcomes such as trust, well-being, and skill enhancement, factors often overlooked in conventional manufacturing metrics. The reviewed literature indicates that Industry 5.0 does not replace human workers but rather positions them at the heart of intelligent production systems, promoting cooperation, empathy, and adaptability as core characteristics of human-centric manufacturing.

The reviewed literature reveals an asymmetry in paradigm maturity: while HMI and HRC are comparatively well-documented with empirical validation, HMC and HCPS frameworks remain largely conceptual or confined to laboratory demonstrations. This gap is significant because these latter paradigms, those involving the highest degree of human cognitive involvement, are precisely the ones most central to Industry 5.0’s human-centric ambitions. The paradox thus exposed highlights a systemic research priority misalignment that future studies must urgently address. Moreover, the evolution of collaboration paradigms is not merely technological but fundamentally socio-technical, reflecting a shift toward systems that emphasize adaptability, trust, and human agency over pure automation efficiency. Yet, inconsistencies across studies in measuring these outcomes underscore the need for standardized evaluation metrics to ensure comparability and practical implementation.

3.3. Synthesis of implementation challenges and research gaps

Table 5 presents a synthesized overview of the key implementation challenges and associated research gaps affecting the adoption of human-machine collaboration in Industry 5.0 manufacturing systems. The table consolidates findings from the thematic analysis by categorizing recurring technological, organizational, ethical, and economic barriers and linking them to their implications for human-centric manufacturing. By explicitly mapping these challenges to unresolved research gaps.

The table demonstrates that realizing human-centric manufacturing extends beyond mere technological maturity. Effective implementation requires the simultaneous consideration of organizational structures, ethical concerns, socio-technical interactions, and economic factors. Addressing these dimensions in isolation limits scalability and practical impact. Consequently, structured implementation frameworks, comprehensive evaluation metrics, and strategies aligned with regulatory and policy objectives are essential. These coordinated approaches help bridge existing gaps, support context-sensitive deployment, and ensure that Industry 5.0 human-machine collaboration delivers sustainable, inclusive, and scalable value across diverse manufacturing environments.

3.4. Implication of the study

This systematic review presents several key implications that transform the synthesized evidence from Tables 2-4 into actionable and theoretical advancements for human-centric manufacturing in Industry 5.0. The implications outlined below emphasize the study’s contributions to bridging research fragmentation and facilitating real-world transitions.

1) Theoretical implication (Advancement of human-centric Industry 5.0 frameworks): By integrating sixty peer-reviewed studies into comprehensive taxonomies (Tables 2-4), this review moves beyond previous fragmented analyses to propose a cohesive theoretical framework. The classification of technological enablers into five clusters clarifies their collective roles and supports the development of dynamic models that capture human-machine synergies. This enriches the conceptual foundations of Industry 5.0 literature and enables future hypothesis-driven investigations into co-evolutionary human-system interactions.

2) Practical/managerial implications (Guidance for industrial implementation): Practitioners and manufacturing managers receive evidence-based directives for designing and implementing human-machine collaboration systems. The mapping of technological clusters to paradigms (discussed further in Section 3.5) provides decision-makers with the tools to select context-appropriate integrations, such as combining collaborative robotics with advanced sensing for human-robot collaboration in assembly lines. These integrations enhance productivity, reduce ergonomic risks, and improve worker well-being. This structured guidance minimizes implementation uncertainties, encourages informed investment in enabling technologies, and supports the scalable adoption of smart manufacturing practices tailored to operational needs.

3) Policy implication (Support for workforce development and governance strategies): The synthesized challenges and gaps identified in Table 4 offer policymakers critical insights for developing supportive frameworks. The review emphasizes the need for targeted workforce upskilling programs, ethical AI governance protocols, and inclusive technology diffusion policies, particularly to address disparities in developing economies. By underscoring the importance of aligning regulations with human-centric outcomes (e.g., trust and safety), this study informs national and international strategies for sustainable and resilient deployment of Industry 5.0 while mitigating socio-ethical risks.

3.5. Interconnections and practical insights

The interaction among technological clusters creates a progressive chain that supports the evolution of collaboration paradigms. Notably, the synergy between the Artificial Intelligence and intelligent algorithms cluster and Digital Twin and Human Digital Twin technologies drives a transition from basic HMI paradigms, which focus on simple monitoring in control rooms, to advanced HMC and HCPS. AI-driven emotion recognition and adaptive task allocation, enhanced by HDTs for real-time cognitive workload simulation, improve transparency and trust. This shift elevates human roles from passive operators to active decision-makers and collaborative partners. This progression is particularly evident in applications like predictive maintenance and adaptive production lines, where these integrations lead to measurable human-centric improvements in decision quality and worker well-being, while fostering reciprocal learning that is often lacking in isolated technology deployments.

Additionally, collaborative robotics combined with sensing, interface, and connectivity technologies (such as AR/VR and wearable sensors) accelerates the development of HRC paradigms. Force-sensing mechanisms and context-aware interfaces facilitate safe co-execution in shared workspaces for assembly and maintenance tasks, resulting in reduced physical workload and improved safety outcomes. The architectures of cyber-physical and human-cyber-physical systems orchestrate these elements in a holistic manner, ensuring systemic resilience across paradigms.

However, these interconnections are not without practical implementation challenges, which reveal deeper insights into adoption barriers. While theoretical synergies suggest seamless progression, real-world constraints, such as the complexity of integrating with legacy systems and the scarcity of data for training reliable human-aware models, often hinder translation, particularly in small and medium enterprises (SMEs) and developing economies. Limited empirical and longitudinal validation, as highlighted in Table 4, presents a significant critique: many documented benefits (e.g., enhanced trust through AI and Digital Twins in HMC/HCPS) arise from controlled laboratory settings, with little evidence of scalability in resource-constrained environments where infrastructure deficits and skill gaps amplify ethical concerns like algorithmic bias and privacy erosion. Therefore, without modular architectures and targeted upskilling, these technological enablers risk exacerbating rather than alleviating human-centric disparities.