Performance-Oriented Digital Twin Technologies for Digital Characters and Avatars: A Bibliometric and Thematic Analysis

Mauricio Hincapié Montoya^a, Jesús Alejandro Guzmán-Ramírez^b, and Alejandro Valencia-Arias^{c, d}

^a Escuela de Artes y Humanidades, Universidad EAFIT, Colombia, 55420

^b Escuela de Artes y Humanidades, Universidad EAFIT, Colombia, 50022

^c Vicerrectoría de Investigación y postgrado, Universidad de Los Lagos, Chile, 5290000

^d Coordinación de Investigación, Corporación Universitaria Americana, Colombia, 50012

maurhin@eafit.edu.co, jaguzman@eafit.edu.co, javalenciar@gmail.com

ABSTRACT

KEYWORDS

1. Introduction

Digital twin technologies have emerged as a major innovation in the realm of digital simulation and modeling. A digital twin can be defined as a virtual replica of a real object, system, or process. The objective of a digital twin is to enable the emulation of its operation, the formulation of predictions, and the enhancement of its performance. In the domain of digital character and avatar development, these technologies enable the progression of avatars from static representations to dynamic entities capable of responding with enhanced precision and personalization to user actions. In this context, performance is defined as the capacity of avatars to effectively and credibly interact in virtual environments, such as video games, virtual reality simulations, or animation platforms. Recent research has validated the capacity of these technologies to accurately replicate behaviors and reactions, emphasizing the utilization of digital twins in simulating the behavior of complex systems, including avatars in virtual environments (Stary, 2023).

These technologies have had a transformative impact on various industries, particularly the video gaming sector, where avatars play a pivotal role in facilitating an immersive experience. Advancements in digital twins have had a notable impact on animation and virtual reality, enabling the creation of more organic and natural characters. The characters undergo a process of real-time adaptation, thereby fostering an enhanced emotional connection with the users. This enhancement in user engagement can be attributed to the integration of spatial, temporal, and real-time data in digital twins, which surpass conventional models and optimize human interaction (Tan et al., 2024). Furthermore, the incorporation of artificial intelligence into avatars enhances their capacity to learn and adapt, thereby ensuring more personalized and realistic experiences (Chaudhary et al., 2024).

A comprehensive understanding of the interplay between digital twin technologies and the efficacy of digital persona creation is imperative. The integration of these systems has been demonstrated to improve the precision, customization, and engagement of digital experiences. Recent research indicates that digital twins have the potential to enhance the design and operation of applications in the metaverse. This enhancement can be achieved by ensuring that avatars accurately reflect the user, thereby facilitating adaptive interactions (Stary, 2023; Wallace, 2023).

The confluence of emerging technologies, including artificial intelligence, the metaverse, and digital twins, unveils novel possibilities in the creation and training of digital personas, particularly in industrial sectors where the training and maintenance of complex products stand to benefit from these innovations (Bordegoni & Ferrise, 2023). Nevertheless, a conspicuous absence of literature persists in extant publications concerning digital twins and avatar creation. This absence is marked by an absence of a discernible framework that would facilitate a synthesis of these two fields. Despite the numerous studies conducted, a coherent conceptual framework remains elusive, hindering the identification and consolidation of emerging trends. The dearth of a unifying framework has impeded the development of theories that address the interactions between avatar performance and digital twin technologies. The dissemination of research across various disciplines, in addition to the absence of consensus on methodologies, further exacerbates the fragmentation of knowledge. The utilization of a multitude of methodologies in the examination of digital twins and avatars impedes the formation of a consolidated comprehension of the phenomenon.

This state of affairs complicates the formulation of models that integrate the various perspectives and dimensions of the problem, thereby impeding progress in the field and hindering the creation of solutions applicable in practical contexts, such as video games, animation, and virtual reality simulations (Jamshidi et al., 2023; Tan et al., 2024; Vanderhorst et al., 2024; Yang et al., 2024).

To achieve the stated objective, there are a series of questions that have guided the development of this research:

1. What is the number of publications per year related to the intersection between performance and digital twin technologies in character and avatar creation?

2. What are the most influential authors and journals in this field?

3. How is research on this topic distributed globally?

4. What are the emerging, growing, and declining keywords in articles on this technological intersection?

The distinguishing characteristic of this study is its emphasis on the development of a scientific framework concerning digital twin technologies and the performance of digital characters. The objective of this study is to systematically analyze existing publications, identify emerging trends, and detect knowledge gaps. This comprehensive approach is expected to update the current state of the art in this rapidly evolving field. The objective of this synthesis is to develop a research agenda that promotes academic and technical progress, thereby consolidating a clearer scientific body of knowledge. The proposed research agenda is informed by a multidisciplinary approach, encompassing domains such as engineering, animation, artificial intelligence, and human-computer interaction. This comprehensive approach endeavors to cultivate a more profound and nuanced comprehension of the nexus between these technologies.

Notwithstanding the proliferation of research on digital twin technologies and digital avatars, several specific gaps remain unaddressed. Existing studies have examined digital twins, avatar design, and artificial intelligence as largely independent research streams, with limited attention to their combined impact on performance-oriented digital actors. Furthermore, previous reviews seldom offer a systematic mapping of the evolution of performance-related concepts over time within this domain. Additionally, they frequently lack an integrated bibliometric perspective that connects technological, performative, and interactional dimensions. The absence of conceptual integration and longitudinal analysis hinders both theoretical consolidation and practical applicability. This underscores the necessity for a focused bibliometric study that explicitly examines the convergence of digital twin technologies and avatar performance.

2. Methodology

This research employs an exploratory analysis based on secondary research sources, in accordance with the parameters set forth in the PRISMA-2020 declaration (Page et al, 2021). This methodology enables an exhaustive bibliometric analysis of literature pertaining to digital actors.

2.1. Search Strategy and Eligibility Criteria

This study was designed as a structured bibliometric investigation aimed at systematically examining the scientific production situated at the intersection between performance-oriented systems and digital twin technologies in the development of digital characters and avatars. The methodological design combines systematic identification procedures with quantitative bibliometric analysis. To ensure transparency in the identification and screening stages, the reporting of the selection process follows the PRISMA 2020 statement (Page et al., 2021). It is important to clarify that PRISMA is adopted strictly as a reporting framework to document how records were identified, screened, and included. The bibliometric analysis itself does not rely on PRISMA as a methodological protocol, but rather uses it to ensure clarity, traceability, and replicability in describing the corpus construction process.

Information Sources

The bibliographic records analyzed in this study were retrieved exclusively from two internationally recognized indexing platforms: Scopus (Elsevier) and Web of Science Core Collection (Clarivate Analytics). These databases were selected due to their structured metadata architecture, multidisciplinary coverage, citation indexing reliability, and established use in bibliometric research (Culbert et al., 2025). Both platforms provide standardized export formats and citation-linked records, which are necessary for performance measurement, co-occurrence mapping, and network visualization.

The searches were executed using the official web-based interfaces of Scopus and Web of Science Core Collection. The final search was conducted on October 18, 2023. The search was not supplemented by additional repositories, grey literature sources, citation chaining, reference list screening, institutional archives, or conference databases. This delimitation was intentional in order to preserve comparability within curated, peer-reviewed, indexed scientific production and to avoid heterogeneity in metadata standards.

Search Strategy

The search strategy was designed in a structured and database-specific manner, taking into account the technical indexing characteristics and syntactic requirements of Scopus and Web of Science. The primary objective was to retrieve literature explicitly addressing the intersection between performance-related concepts and technologies associated with digital twins and digital actors, while ensuring thematic coherence and minimizing false positives.

The development of the search equations was grounded in three clearly defined conceptual axes: (i) terms related to acting and performance, (ii) concepts associated with digital twin technologies and motion capture (mocap) systems, and (iii) the intersection of these dimensions within specific indexed fields, namely titles and author keywords. These terms were systematically combined using the Boolean operators AND and OR in order to maximize conceptual coverage without compromising semantic precision.

The temporal window was established from 1995 to 2023. This period was selected to capture the earliest documented developments related to performance-based digital representations, as well as their subsequent evolution in the context of digital twin technologies, artificial intelligence, and immersive virtual environments. This timeframe enables a longitudinal perspective on the thematic development and consolidation of the field.

The exact search equations implemented in each database were as follows:

For the Scopus database:

((TITLE(«Acting» OR «Actors» OR «Performers» OR «Performer» OR «Performing» OR «Performance» OR «Dramaturgy») AND TITLE(«digital twin» OR «mocap characters» OR «digital personas» OR «mocap actors» OR «Digital actors» OR «mocap avatars» OR «mocap beings»)) OR (AUTHKEY («Acting» OR «Actors» OR «Performers» OR «Performer» OR «Performing» OR «Performance» OR «Dramaturgy») AND AUTHKEY(«Mocap people» OR «Mocap characters» OR «Mocap personas» OR «Digital twin» OR «Digital actors» OR «Mocap avatars» OR «mocap beings»)))

For the Web of Science database:

((TI=(«Acting» OR «Actors» OR «Performers» OR «Performer» OR «Performing» OR «Performance» OR «Dramaturgy») AND TI=(«digital twin» OR «mocap characters» OR «digital personas» OR «mocap actors» OR «Digital actors» OR «mocap avatars» OR «mocap beings» )) OR (AK= («Acting» OR «Actors» OR «Performers» OR « Performer» OR «Performing» OR «Performance» OR «Dramaturgy») AND AK=(«Mocap people» OR «Mocap characters» OR «Mocap personas» OR «Digital twin» OR «Digital actors» OR «Mocap avatars» OR «mocap beings»)))

These equations were executed directly within the official interfaces of each database. No initial language restrictions were applied during the search phase; however, thematic relevance was subsequently verified during the screening process.

The construction of the search equations followed a balance between sensitivity and specificity. The inclusion of lexical variations and synonymous expressions enabled the capture of diverse terminological usages within the literature. At the same time, restricting the search to structured indexing fields (TITLE, AUTHKEY, TI, and AK) ensured thematic alignment with the study’s focus and reduced retrieval noise.

By explicitly presenting the exact equations applied in each database, this strategy guarantees methodological transparency, traceability, and replicability of the search process.

Eligibility Criteria

Eligibility criteria were defined a priori to ensure operational clarity and replicability. A record was considered eligible if it satisfied all of the following conditions:

First, it had to be indexed in either Scopus or Web of Science Core Collection. Second, it had to be classified as a peer-reviewed journal article. Third, it had to fall within the defined temporal window. Fourth, it had to be written in English. Fifth, it had to substantively address digital twin technologies in relation to performance systems, simulation environments, digital character creation, avatar modeling, immersive systems, or virtual representation frameworks.

The exclusion process was applied in three sequential phases. The first phase consisted of duplicate removal and elimination of non-article document types. Duplicates were identified using DOI comparison as the primary criterion, supplemented by title and author matching when necessary. The second phase excluded records with incomplete essential bibliographic metadata. In this study, «incomplete indexing» refers specifically to records lacking critical metadata required for bibliometric processing, such as missing author information, publication year, journal title, or DOI. The third phase involved thematic screening based on title and abstract review. Records were excluded as «out of scope» if the central focus did not substantively align with the intersection between digital twin technologies and performance-oriented digital systems.

These criteria were operationalized to ensure that another researcher applying identical parameters could reproduce the same corpus under the same conditions.

2.2. Selection Process and PRISMA Reporting Flow

The selection workflow followed a structured and transparent sequence, documented according to the PRISMA 2020 reporting framework (Page et al., 2021). The use of PRISMA in this study serves exclusively to provide a clear account of how records progressed from initial identification to final inclusion. It does not constitute a methodological evaluation of empirical outcomes, as the present study is bibliometric in nature.

Following retrieval from Scopus and Web of Science, all records were exported in standardized bibliographic formats and consolidated into a single master dataset. The identification stage yielded the total number of records retrieved from both databases. Automatic duplicate detection was performed through DOI matching and exact title comparison. This procedure was followed by manual verification to ensure that variations in metadata formatting did not result in incorrect duplicate classification. No record was excluded solely through automated filtering without subsequent human validation.

Title and abstract screening was conducted independently by two researchers. Each reviewer assessed thematic alignment with the predefined eligibility criteria. Where discrepancies arose, records were discussed jointly until consensus was reached. This independent screening procedure enhanced reliability and reduced subjective bias in inclusion decisions.

Full metadata review was subsequently conducted to confirm alignment with the eligibility criteria and verify the completeness of bibliographic information. Records that did not meet the operational definitions of eligibility were excluded during this stage.

The complete reduction pathway—from initial identification to final inclusion—is presented in Figure 1. The PRISMA flow diagram (Figure 1) summarizes the number of records identified, duplicates removed, records screened, exclusions applied at each stage, and the final corpus size of 131 articles. The narrative description of this progression is intentionally concentrated in this subsection to avoid redundancy elsewhere in the methodology.

Figure 1 illustrates the identification, screening, and selection process of the studies included in the bibliometric analysis. The diagram details the initial retrieval of records from Scopus and Web of Science, the removal of duplicate entries, the sequential application of exclusion criteria, and the final composition of the analytical corpus consisting of 131 articles. This procedure ensures methodological transparency, rigor, and reproducibility.

Automation tools were used exclusively for structured data management tasks, such as duplicate detection and metadata filtering. The Excel-based toolkit applied in this study facilitated record organization and cross-validation but did not autonomously exclude records. All inclusion and exclusion decisions were confirmed by human reviewers.

2.3. Data Preparation and Bibliometric Analysis

Once the final corpus of 131 articles was established, a systematic data preparation workflow was implemented to ensure analytical consistency, transparency, and reproducibility. Bibliographic metadata were exported from both databases and merged into a unified dataset.

The extracted metadata fields included authors, article title, year of publication, journal/source title, author affiliations, author-provided keywords, indexed keywords, citation counts, DOI, and document type. These fields were selected because they form the basis for performance measurement, relational mapping, and thematic evolution analysis.

Data extraction was conducted independently by two researchers. The extracted datasets were cross-compared to identify inconsistencies or discrepancies. Where differences were detected, records were manually verified against the original database entries. This dual-verification approach ensured accuracy and minimized transcription errors.

Database merging was conducted using DOI as the primary matching variable. In cases where DOI was unavailable, title and author matching were applied. Standardization procedures were implemented to harmonize metadata across platforms. These procedures included normalization of author names, consolidation of keyword variants (e. g., singular/plural harmonization), removal of encoding inconsistencies, and correction of formatting discrepancies.

No statistical imputation was performed for missing values. Records lacking essential metadata required for analysis had been excluded during the eligibility phase.

The bibliometric analysis comprised three interrelated indicator domains: performance indicators, structural indicators, and temporal indicators.

Performance indicators included annual publication trends, cumulative citation counts, average citations per article, identification of the most productive authors, most influential journals, and geographical distribution of scientific production. These metrics were calculated through frequency tabulation and descriptive aggregation using Microsoft Excel.

Structural indicators focused on relational mapping of conceptual patterns. Keyword co-occurrence networks were generated using VOSviewer to identify thematic clusters and conceptual proximities. Prior to network construction, keywords were standardized to reduce artificial fragmentation of semantically equivalent terms. Threshold values for minimum keyword occurrence were defined to ensure meaningful cluster detection. A minimum occurrence threshold of three was applied in VOSviewer to filter keywords and construct the co-occurrence network, allowing the identification of significant thematic clusters while avoiding excessive fragmentation of the network.

Temporal indicators were constructed by dividing the study period into defined intervals and analyzing shifts in keyword prominence across phases. This temporal segmentation enabled identification of thematic transitions, emergent research fronts, and evolving conceptual configurations.

In this study, the term «synthesis» refers to structured aggregation, relational mapping, and visualization of bibliographic metadata. It does not refer to statistical meta-analysis of empirical effect sizes. The analytical workflow proceeded sequentially from extraction to cleaning, standardization, aggregation, mapping, and visualization, forming a continuous and reproducible process.

2.4. Bias, Limitations, and Robustness

Because this study is bibliometric, the concept of reporting bias differs from its conventional meaning in systematic reviews of empirical outcomes. In this context, potential sources of bias are primarily associated with database coverage, indexing practices, and metadata completeness.

First, database coverage limitations may exclude relevant publications not indexed in Scopus or Web of Science. Second, terminological variation may result in incomplete retrieval if certain synonyms were not captured by the search strings. Third, language restrictions (English-only inclusion) may limit representativeness. Fourth, exclusion of conference proceedings and grey literature may affect coverage in rapidly evolving technological domains. Fifth, incomplete metadata may introduce structural gaps in mapping analyses.

These limitations were mitigated through transparent search design, dual-database retrieval, structured duplicate verification, and explicit eligibility criteria. Nevertheless, they represent inherent constraints of bibliometric research based on indexed journal articles.

The concept of «certainty of evidence» in this study refers to the robustness and reproducibility of the bibliometric dataset and analytical workflow rather than to the empirical validity of findings within the included articles. Robustness is supported by explicit search strings, clearly defined eligibility parameters, independent screening procedures, structured duplicate validation, standardized data extraction protocols, and transparent indicator computation methods.

The dataset is fully traceable to indexed bibliographic records. Under identical search parameters and eligibility criteria, the workflow can be replicated. This methodological transparency constitutes the principal basis for confidence in the analytical process.

3. Results

The results of the current bibliometric analysis are significant, as shown in Figure 2, which also shows a notable cubic growth in the production of articles on digital actors. In the years 2020, 2021, 2022, and 2023, the increase in interest and attention on the subject of digital actors has been particularly notable, with an 88.36 % increase observed. These years have been distinguished by the observation of a greater number of scientific publications on digital actors, indicating a constantly expanding research trend in this particular field.

All descriptive statements presented in this section are derived directly from the bibliometric metadata of the 131 selected records and correspond to measurable indicators such as document counts, citation frequencies, keyword co-occurrence strength, and temporal distribution. Interpretations are therefore anchored in corpus-based quantitative evidence.

Temporal evolution of scientific publications related to digital actors and digital twin technologies from 1995 to 2023. The figure reveals a sustained and accelerated growth in research output, with a particularly pronounced increase after 2020, reflecting the consolidation of the field and the growing academic interest driven by advances in artificial intelligence, simulation, and immersive digital environments.

Beyond isolated yearly examples, the temporal evolution of the field reveals three distinguishable phases. The first phase (1995–2009) is characterized by sporadic publication activity, reflecting an exploratory stage in which digital representation and virtual performance concepts were emerging. The second phase (2010–2019) demonstrates gradual consolidation, with a steady increase in publications associated with the integration of digital twin architectures and simulation-based systems. The third phase (2020–2023) represents an acceleration stage, accounting for a substantial proportion of the total corpus and reflecting the convergence of artificial intelligence, smart manufacturing, and performance-driven digital twin applications. This period-based interpretation offers a more comprehensive understanding of field maturation than referencing individual years alone.

The analysis of lead authors identified three distinct groups, as shown in Figure 3. The first group comprises leading productivity and impact experts, including S. Wang. These researchers have demonstrated outstanding scientific production and considerable impact in terms of citations. The second group stands out for high influence despite a lower volume of scientific publication. These writers include Y. Li, B. Lei, L. Li, and C. Mao. Finally, the third group is characterized mainly by its scientific productivity, despite not always being the most cited; authors such as X. Song and X. He are notable examples.

To provide explicit quantitative support for the interpretation of author influence, productivity within the 131-document corpus was calculated based on the number of indexed publications per author, while impact was assessed using citation counts derived from the same metadata source. The most productive authors are those contributing the highest number of documents within the analyzed dataset, whereas highly cited authors are those whose publications accumulate the greatest citation totals. The classification presented in Figure 3 is therefore grounded in corpus-derived indicators rather than narrative interpretation. This clarification ensures that statements regarding productivity and influence are directly supported by measurable bibliometric evidence.

Network of leading authors in the research field based on productivity and citation impact indicators. The visualization identifies three distinct author profiles: highly productive and highly cited researchers, authors with lower publication volume but high scientific influence, and authors characterized by high productivity with comparatively lower citation impact. This structure highlights the intellectual leadership and collaboration patterns within digital actor research.

Figure 4 depicts the findings of the bibliometric analysis of the key journals in the field of digital actors. The analysis revealed three distinct groups of journals. The first group comprises the most prolific and influential journals, exemplified by the Journal of Manufacturing Systems. These journals have demonstrated exceptional scientific production and a considerable impact in terms of citations. The second group includes journals such as the Journal of Industrial Information Integration and IEEE Computer Graphics and Applications, which have a lower scientific productivity index but a notable impact. The third group of journals, which includes Applied Sciences and Jusianji Jicheng Zhizao Xitong/Computer Integrated Manufacturing Systems, CIMS, stands out mainly for scientific productivity, although not necessarily for high citation counts.

Distribution and clustering of the main scientific journals publishing research on digital actors and digital twin technologies. The figure distinguishes journals with high productivity and high impact, journals with lower publication volume but strong citation influence, and outlets characterized by substantial output with moderate citation levels. This distribution underscores the interdisciplinary nature of the field, spanning manufacturing systems, industrial informatics, and computer graphics.

Journal influence was assessed using two primary indicators: number of documents published within the corpus and total citation accumulation. The characterization of journals as «prolific» or «influential» therefore corresponds to their measurable contribution in terms of publication frequency and citation impact within the 131 analyzed studies. This quantitative grounding reinforces the interpretation of disciplinary concentration and avoids unsupported evaluative language.

In this bibliometric study on digital actors, the analysis by countries identified three groups. The first group comprises the leading countries in both productivity and impact, as shown in Figure 5, with China standing out for its outstanding scientific production and considerable impact in terms of citations. The second group which includes the United States, Switzerland, and Italy, shows strong citation impact despite having a lower scientific productivity index. Finally, the third group is characterized primarily by high scientific productivity, although not consistently by high citation counts; Germany is the most notable example. This country-level analysis offers a comprehensive overview of the various aspects of each nation's contribution to the field of digital players.

Geographic distribution of scientific production by country in the field of digital actors. The figure highlights countries leading in both publication volume and citation impact, as well as regions with relevant but differentiated contributions in terms of productivity and visibility. These patterns reflect the influence of national research policies, technological investment, and innovation ecosystems on the development of the field.

Country-level analysis was similarly derived from aggregated publication counts and citation totals associated with institutional affiliations in the dataset. Leading countries are those demonstrating both high document production and sustained citation impact within the corpus. This approach ensures that geographic interpretations are based on verifiable bibliometric indicators rather than descriptive generalizations.

The current bibliometric research examined the thematic development in literature related to digital actors during the period spanning 1995 to 2023, as shown in Figure 6. This analysis reveals that the emergence of pivotal terms such as «virtual reality» in 1995 (Heim, 1995) served as a pivotal point and marked a significant beginning in the study of digital actors. The thematic evolution analysis was conducted using author keywords and indexed terms extracted from the 131 selected records. Within this dataset, the earliest occurrence of the keyword «virtual reality» appears in publications indexed in 1995, marking one of the initial conceptual anchors of the field. This observation is directly traceable to the analyzed corpus and supports the interpretation that early research attention focused on immersive digital environments before progressively incorporating digital twin technologies and performance optimization concepts. The evolution of topics of interest and the trajectory of research on digital actors can be elucidated by examining the prevalence of emerging topics over time, including «Wearable Devices», «Performance of Production», «Surrogate Model», and «Machine Learning».

Thematic evolution of key concepts associated with digital actors over the period 1995–2023. The figure illustrates the progressive transition from early research focused on virtual reality toward more advanced themes such as wearable devices, production performance, surrogate models, and machine learning. This evolution evidences the conceptual maturation of the field and its increasing orientation toward intelligent and performance-driven systems.

Figure 7 presents a graphical representation of the primary keyword co-occurrence network, which is divided into a total of nine thematic groups. The group highlighted in red stands out from the others due to the inclusion of words such as «artificial intelligence», «machine learning», «building performance simulation», and «data-driven model», which collectively suggest that research has focused on these topics. The subsequent group is dark green and includes terms such as «smart manufacturing», «predictive maintenance», «automation», «control», «optimization», and «delays». This group indicates a conceptual affinity to various aspects of manufacturing and process management. The other color groups (light blue, dark blue, yellow, light purple, brown, pink, and orange) also provide a complete picture of the variety of themes and conceptual connections found in the literature on digital actors.

Keyword co-occurrence network depicting the thematic structure of research on digital actors. The identified clusters group concepts related to artificial intelligence, machine learning, performance simulation, smart manufacturing, and data-driven models. Node density and proximity indicate conceptual affinity, enabling the identification of dominant research cores as well as emerging thematic areas.

A Cartesian plane has been proposed to represent the relationship between keyword usage frequency on the X-axis and usage validity on the Y-axis in the context of this bibliometric research on digital actors, as illustrated in Figure 8. This method enables the identification of four distinct quadrants based on these two dimensions. In Quadrant 2, keywords such as «Artificial Intelligence», «Surrogate Model», «Smart Manufacturing», and «Deep Learning» are observed, which are relatively uncommon in the literature but highly current, indicating that they are emerging terms. This growing interest in the field is evident. In Quadrant 1, however, there are terms such as «Industry 4.0», «Simulation», «Internet of Things», and «Machine Learning» that are common and up to date, indicating that they are well-established concepts that are continually being expanded in the investigation.

Cartesian plane relating keyword usage frequency (X-axis) to conceptual validity or recency (Y-axis). This representation enables the classification of keywords into four quadrants, distinguishing well-established and frequently used concepts from emerging terms characterized by high current relevance but lower frequency. The analysis supports the identification of research trends and future research opportunities in the digital actor domain.

4. Discussion

This section examines the growing body of scientific literature on digital actors in light of the results obtained in terms of bibliometric indicators. It discusses the practical implications of the findings, acknowledges the limitations of the study, and highlights major research gaps, thereby providing a valuable research agenda for future studies on digital actors. Additionally, keywords are classified according to their function.

4.1. Analysis of the Growth of Scientific Literature on Digital Actors

The growth of scientific production in this field is not inferred from isolated examples but from the annual publication counts derived from the 131-document corpus. Between 1995 and 2009, publication activity remained limited and sporadic, reflecting an exploratory phase. A gradual increase is observed between 2010 and 2019, indicating thematic consolidation. The most significant expansion occurs from 2020 to 2023, when the annual number of publications increases markedly, accounting for a substantial proportion of the total corpus. This quantitative distribution confirms a clear acceleration pattern in recent years and demonstrates that the observed growth is supported by measurable bibliometric data rather than narrative illustration.

In particular, the year 2020 marked a foundational stage in the consolidation of digital twin–based approaches, with comprehensive reviews outlining key concepts, enabling technologies, and industrial applications of digital twins, thereby providing a theoretical and methodological basis for subsequent research on digital actors (Liu et al., 2020).

In 2021, a study introduced a generic methodology and a digital twin framework for mapping performance in the context of zero defect manufacturing (ZDM), proposing a structured approach to linking performance indicators with design and production processes (Psarommatis, 2021). This contribution illustrates the use of digital twin–based representations to support manufacturing-related performance assessment.

In 2022, research further consolidated the application of digital twin technologies within smart manufacturing environments, emphasizing performance optimization and system integration. For instance, Li et al. (2022) analyzed the implementation of digital twin architectures to enhance production monitoring, data synchronization, and operational efficiency in industrial contexts. This contribution illustrates how digital representations were increasingly used to support performance management and real-time decision-making processes.

In 2023, a digital twin framework was proposed for the automated detection of fault sources and the prediction of comfort performance in non-residential buildings in Norway. This study reported the use of digital twins to monitor and assess building performance using real operational data (Hosamo et al., 2023).

4.2. Analysis of Research References on Digital Actors

The findings indicate that research on digital actors is theoretically structured around three main axes: energy and performance simulation, control and realism of digital actors, and the application of digital twins in industrial and built environments. The prominence of simulation-based studies highlights the growing recognition of digital actors as analytical tools capable of supporting optimization and sustainability objectives, particularly in energy-intensive systems such as electric mobility and manufacturing (Liu et al., 2020; Li et al., 2022).

From a theoretical standpoint, the contributions of Y. Li emphasize the importance of unified planning and control frameworks for autonomous digital actors, reinforcing the role of motion planning and control theory in advancing adaptability and versatility in virtual agents (Li & Li, 2008). In parallel, studies focusing on advanced digital technologies expand this perspective by integrating data-driven and intelligent approaches, which strengthen the functional realism and decision-making capacity of digital actors (He et al., 2021; Wang et al., 2022).

The concentration of influential publications in journals such as Journal of Manufacturing Systems, Journal of Industrial Information Integration, and IEEE Computer Graphics and Applications reflects the interdisciplinary nature of the field. Manufacturing-oriented studies demonstrate how digital twins contribute to process optimization and quality improvement, supporting theoretical models that link virtual representations with physical system performance (Zhou et al., 2021; Li et al., 2022). At the same time, research disseminated through computer graphics venues, exemplified by the Digital Emily project, advances theoretical discussions on realism, human likeness, and fidelity in digital actors, thereby extending the field toward human-centered and perceptual dimensions (Alexander et al., 2010).

The role of Applied Sciences further underscores the theoretical relevance of interoperability between different digital modeling paradigms, such as BIM and BEM. This integration supports a systems-based theoretical view in which digital actors function as connectors across disciplinary boundaries, enabling holistic analysis and decision support in complex environments (Bastos Porsani et al., 2021).

At the geographical level, China’s leadership suggests a strong alignment between theoretical development and industrial application, particularly in the context of smart manufacturing and digital twins (Li et al., 2022). In contrast, the influence of the United States, driven by high-fidelity digital human modeling, reflects a theoretical emphasis on realism, visualization, and human–machine interaction (Alexander et al., 2010). European contributions provide complementary theoretical insights, focusing on methodological rigor and process evaluation. For example, Switzerland emphasizes generic methodologies for defect-free manufacturing (Psarommatis, 2021), Italy advances performance evaluation frameworks for production lines (Fera et al., 2019), and Germany integrates augmented reality with digital twins for process monitoring, reinforcing theories of cyber–physical integration (Yi et al., 2021).

Overall, the discussion confirms that the theoretical evolution of digital actors is driven by the convergence of simulation science, control theory, computer graphics, and industrial engineering. This convergence positions digital actors not merely as virtual representations, but as central theoretical constructs for understanding, optimizing, and transforming complex digital–physical systems across multiple application domains.

4.3. Analysis of the Thematic Evolution of Digital Actors

The reviewed studies reinforce the theoretical understanding of digital actors as dynamic entities that mediate between physical systems and their digital representations through continuous data exchange and intelligent processing. The integration of wearable devices into digital twin frameworks extends this conceptualization by embedding human factors such as operator health, ergonomics, and performance directly into the digital actor ecosystem. From a theoretical perspective, this integration supports socio-technical system theories, in which humans, machines, and digital models are treated as interdependent components of a unified production system (Fera et al., 2019).

Production performance emerges as a central theoretical construct in the evolution of digital actors. The use of factory-level digital twins to evaluate the impact of wireless communication highlights how performance is no longer viewed solely as an output metric, but as a dynamic variable influenced by connectivity, data latency, and system responsiveness. This perspective aligns with cyber–physical systems theory, where digital twins function as analytical layers that enable real-time monitoring, prediction, and optimization of manufacturing processes, thereby enhancing competitiveness and operational efficiency (Mogensen et al., 2021).

The incorporation of surrogate models into digital twin development represents a significant theoretical advancement in managing system complexity. Multi-level fusion modeling approaches enable digital actors to capture both geometric and performance-related characteristics with high fidelity, while reducing computational cost. Theoretically, surrogate modeling supports abstraction and approximation principles, allowing digital twins to balance accuracy and efficiency. This capability strengthens the role of digital actors as decision-support mechanisms that can operate effectively under real-time constraints (He et al., 2022).

The growing integration of machine learning into digital twin architectures further reshapes the theoretical foundations of digital actors. By enabling predictive capabilities such as forecasting structural performance in applications like battery supports machine learning transforms digital twins from descriptive and diagnostic tools into predictive and prescriptive systems. This evolution aligns with data-driven modeling theories, where learning algorithms enhance adaptability and continuously refine digital representations based on new data streams. As a result, digital actors become active agents in optimization and decision-making processes across industrial and engineering domains (He et al., 2023).

Overall, these contributions suggest that the theoretical trajectory of digital actors is moving toward increasingly intelligent, human-aware, and performance-oriented systems. The convergence of wearable technologies, performance modeling, surrogate-based simulation, and machine learning positions digital actors as core constructs in digital transformation theory, enabling a more holistic understanding of complex production environments and reinforcing their strategic role in advanced manufacturing and industrial innovation.

4.4. Analysis of Thematic Clusters on Digital Actors

The thematic structure of the keyword co-occurrence network reflects the theoretical consolidation of digital actors around data-driven intelligence and predictive modeling. The increasing importance of artificial intelligence, machine learning, and building performance simulation is indicative of a shift toward adaptive systems that are capable of learning from data and supporting anticipatory decision-making. The integration of data-driven models in domains such as bioprocess digital twins and built heritage management underscores the role of digital actors as dynamic system representations that combine semantic, geometric, and performance information, reinforcing holistic and lifecycle-oriented theoretical perspectives (Massafra et al., 2023; Park et al., 2023).

The second thematic focus centers on smart manufacturing, automation, and predictive maintenance, positioning digital actors within the framework of cyber–physical production system theory. Research on process monitoring, open process automation, and digitally enabled network environments demonstrates how digital actors function as control-oriented and network-aware agents that enable closed-loop optimization and real-time performance management. These findings extend the theoretical scope of digital actors toward scalable, interoperable, and intelligent entities capable of operating effectively in highly connected and dynamic environments (Qamsane et al., 2022; Qi et al., 2023).

4.5. Analysis of the Frequency and Conceptual Validity of Studies on Digital Actors

The identification of emerging concepts in Quadrant 2 of the Cartesian plane underscores the theoretical shift of digital actors toward intelligence-driven and adaptive paradigms. The increasing prevalence of artificial intelligence, surrogate models, and smart manufacturing indicates a shift in the conceptualization of digital actors from static digital representations to learning and predictive systems. From a theoretical standpoint, the integration of heterogeneous models and real-time data streams is facilitated by artificial intelligence, thereby enabling the replication of complex systems with high fidelity by digital twins. This, in turn, supports real-time optimization and decision-making processes. This evolution aligns digital actors with theories of intelligent cyber–physical systems, where adaptability, autonomy, and continuous learning are central characteristics (Lai & Chang, 2021).

The employment of surrogate models serves to further substantiate this theoretical trajectory by addressing the challenge of computational complexity that is inherent to high-fidelity simulations. The capacity of surrogate modeling to facilitate expeditious and precise performance evaluation undergirds abstraction and approximation theories. Consequently, digital actors are rendered efficacious analytical proxies for physical systems. In the context of smart manufacturing, the predictive capabilities of surrogate models enhance flexibility and responsiveness, thereby reinforcing the theoretical perspective that digital actors function as enablers of agile and adaptive production systems within dynamic industrial environments (Chua et al., 2022; Guo et al., 2023).

Conversely, the concepts identified in Quadrant 1 reflect the theoretical consolidation and maturity of the field. The prevailing digital actor research is grounded in a series of foundational pillars, including Industry 4.0, simulation, the Internet of Things, and machine learning. The Industry 4.0 paradigm provides the overarching theoretical framework in which digital actors operate, emphasizing connectivity, automation, and data-driven optimization in smart factories. In the context of predictive maintenance scenarios, simulation-based digital twins have been shown to support theories of proactive system management by enabling the anticipation of degradation, wear, and failure in production assets (Longo et al., 2019; Yang et al., 2022).

The integration of the Internet of Things (IoT) with machine learning further strengthens the theoretical foundations of digital actors by enabling continuous data acquisition and advanced analytics. The Internet of Things (IoT) facilitates real-time interaction between physical and digital systems, while machine learning enhances pattern recognition, prediction, and decision support. Consequently, these technologies establish digital actors as pivotal components of intelligent, interconnected ecosystems, demonstrating the capacity to function across industrial and service-oriented domains. This theoretical consolidation suggests that digital actors are evolving into core constructs of digital transformation, combining mature technological foundations with emerging intelligent capabilities to support complex, data-intensive environments.

4.6. Comparative Analysis with Previous Bibliometric Studies

The present study is positioned within a well-established body of literature that has applied bibliometric methods to the analysis of digital twin technologies and related digital systems. However, its contribution differs substantially from previous studies in terms of thematic scope, analytical focus, and interpretative outcomes.

Several prior bibliometric and systematic reviews have examined digital twin technologies from predominantly technical and industrial perspectives. For example, Ofosu et al. (2022) provide a comprehensive systematic review and bibliometric analysis focused on digital twin architectures, enabling technologies, and industrial applications within digital business strategies. Similarly, J. Wang et al. (2024) analyze the conceptual structure and influencing factors of digital twin research, with particular emphasis on manufacturing systems, cyber–physical systems, and industrial performance. While these studies offer a robust overview of the technological evolution and strategic relevance of digital twins, they do not explicitly address their intersection with performance-related dimensions in the creation of digital characters, avatars, or digital actors.

In a related but distinct research stream, Lu et al. (2025) conduct a bibliometric and content analysis of the virtual avatar field, focusing primarily on human–computer interaction, interface design, and user experience. Although this work advances the understanding of avatar-related research, it does not incorporate the conceptual framework of digital twins, nor does it examine the role of simulation-driven or AI-enhanced performance models in avatar development.

Additionally, Boussetta et al. (2024) present a bibliometric analysis of artificial intelligence and digital infrastructure in the context of sustainable industry and corporate performance, emphasizing smart optimization technologies. Despite its relevance to digital transformation and performance optimization, this study remains anchored in organizational and industrial domains and does not consider the performative or representational dimensions of digital entities such as avatars or digital actors.

In contrast, the present study adopts a more integrative and domain-specific perspective by explicitly examining the intersection between digital twin technologies and performance in the creation of digital characters and avatars. Beyond mapping publication trends and influential actors, this research incorporates variables related to simulation, artificial intelligence, machine learning, surrogate models, and adaptive performance. By bridging disciplines such as engineering, digital animation, artificial intelligence, and human–computer interaction, the study moves beyond purely technological or interface-oriented analyses.

Accordingly, the main differential contribution of this work lies in identifying and systematizing an emerging subfield that connects digital twin technologies with performance-oriented digital actors. This bibliometric synthesis not only clarifies the conceptual boundaries of this intersection but also highlights underexplored research gaps and proposes a future research agenda that has not been explicitly addressed in previous bibliometric reviews.

4.7. Classification of Keywords Related to Digital Actors According to Their Function

Table 1 presents a ranking of the most popular and rapidly expanding keywords associated with digital players, organized by feature. This organization makes it simple to identify the core traits and particular uses of each of these categorized features. Due to its ability to analyze trends in the body of scientific literature in great detail, this table is a crucial tool in bibliometric research on digital actors. This classification allows for the observation of how specific keywords have gained importance within industry, which may indicate the emergence of new research areas. Furthermore, a more comprehensive and contextualized understanding of the key concepts utilized in the discourse surrounding digital actors in academia is provided by identifying the characteristics and applications associated with each categorized function. This not only enhances our comprehension of the subject matter but also establishes a robust framework within which to develop novel scientific research questions and future studies.

4.8. Theoretical Implications

The theoretical underpinnings of this approach encompass the analysis of the frequency of publications per year, the predominant theoretical references, the thematic evolution, the co-occurrence of keywords, and the examination of emerging and growing keywords. This comprehensive analysis provides a vision of the current state and future trends in this field of research. This approach facilitates the identification of not only publication patterns but also the identification of areas of greatest interest and development, as well as gaps in knowledge that require greater research attention.

A meticulous examination of the frequency of publications per year discloses a substantial augmentation in scientific production between 2020 and 2023. This increase can be attributed to technological advancements and the growing importance of digitalization in various sectors. The growth in the number of publications in recent years signifies a consolidation and expansion of research in this field. This underscores the significance of emerging technologies such as digital twins and artificial intelligence.

The identification of the principal theoretical references is essential for comprehending the foundational contributions to the evolution of knowledge about digital actors. Notable authors such as S. Wang and B. Lei have demonstrated remarkable influence in terms of both impact and productivity. The scope of their research encompasses a range of subjects, including the energy consumed by electric vehicles and digital twins in smart manufacturing. The researchers' contributions have laid the groundwork for future studies by establishing the theoretical foundations for research in this area. Their work also demonstrates the value of integrating multiple disciplines and methodological approaches when addressing the complex challenges posed by digital actors.

The application of keyword co-occurrence analysis and the identification of emerging and growing keywords permit the mapping of thematic affinities and trends within the field. The primary cluster, comprising terms such as «artificial intelligence», «machine learning», «building performance simulation», and «data-driven model», underscores the pivotal role of artificial intelligence and data-driven models. These concepts reflect the direction in which research is heading, emphasizing the necessity to develop more intelligent and adaptive systems to enhance efficiency and performance in various applications.

The analysis of emerging and growing keywords in quadrant 2, such as «artificial intelligence», «surrogate model», and «smart manufacturing», reveals the growing importance of these technologies in the industrial and manufacturing field. The integration of artificial intelligence and surrogate models has led to a paradigm shift in the design and operation of systems, while smart manufacturing is redefining production processes through the integration of advanced technologies and predictive approaches.

The identification of research gaps is imperative to inform future research initiatives. Despite the considerable progress achieved in domains such as simulation and the Internet of Things (IoT), there remain significant challenges to be addressed, particularly in the integration and harmonization of diverse technologies to develop genuinely intelligent and autonomous systems. A comprehensive investigation into the ethical and social ramifications of the integration of these technologies is imperative. This inquiry must ascertain that their implementation will yield societal benefits that are both equitable and sustainable.

4.9. Practical Implications

The practical implications of this study are numerous and extend to various areas of application. The thematic evolution from the study of virtual reality towards areas such as wearable devices, production performance, replacement models, and machine learning reflects a significant change in the research priorities and approaches. This thematic shift signifies a heightened focus on the incorporation of sophisticated technologies into pragmatic applications. Consequently, this trend fosters enhanced efficiency and customization within diverse industrial and commercial domains.

The thematic clusters demonstrate a robust conceptual congruence among terms such as «artificial intelligence», «machine learning», «building performance simulation», and «data-driven model». This convergence of concepts indicates a growing interconnectedness of the practical applications of these technologies. For instance, the integration of artificial intelligence and machine learning in the simulation of building performance has the potential to enhance the energy efficiency and sustainability of buildings to a considerable extent. In a similar vein, data-driven models have the capacity to facilitate real-time prediction and optimization, a capability that is imperative for the effective management and operation of complex infrastructures.

With respect to the currency and frequency of keywords, concepts such as «artificial intelligence», «substitution model», «smart manufacturing», and «deep learning» emerge as promising research areas. The advent of these technologies portends a paradigm shift in the realm of manufacturing, with the potential to transform production systems into entities that exhibit high levels of adaptability and autonomy. The implementation of substitution models and deep learning has the potential to optimize processes and reduce costs. Smart manufacturing, on the other hand, can improve flexibility and responsiveness to market demands. These innovations offer dual benefits, enhancing competitiveness in the corporate sector while concomitantly fostering environmental sustainability.

Conversely, the emergence of emerging concepts such as «Industry 4.0», «simulation», «Internet of Things», and «machine learning» underscores the enduring significance of these technologies in the ongoing evolution of modern industry. The notion of Industry 4.0, which underscores the digitalization and interconnection of processes, continues to serve as a foundational element for the digital transformation of contemporary industry. The integration of simulation and the Internet of Things facilitates real-time monitoring and control, thereby enhancing efficiency and reducing downtime. Machine learning, defined as the analysis of large volumes of data, has been shown to facilitate informed and accurate decision-making. This capacity is critical for innovation and continuous improvement in various field.

The practical implications extend beyond the transformation of the industrial and manufacturing sectors. In the domain of health care, the integration of technologies such as artificial intelligence and wearable devices has the potential to transform the landscape of medical care. Wearable devices are capable of continuous monitoring of vital parameters, thereby facilitating early disease detection and proactive patient health management. The integration of artificial intelligence with extensive health data has the potential to enhance diagnostic precision, personalize treatment regimens, and optimize clinical outcomes while reducing treatment costs.

In the domain of education, the integration of digital entities presents novel opportunities to enhance the learning experience and administrative efficiency. The utilization of simulations and data-driven models facilitates the development of more interactive and realistic learning environments, thereby promoting the comprehension of complex concepts through practical experiences. Moreover, artificial intelligence has the capacity to personalize educational content and teaching methods according to the individual needs of students, thereby enhancing knowledge retention and academic performance.

In the domain of urban management, the concepts of smart manufacturing and the Internet of Things are foundational for the development of smart cities. The implementation of these technologies has been demonstrated to facilitate the efficient management of urban resources, including energy and water. Moreover, digital twins have the capacity to simulate and optimize the operation of critical infrastructure. This capability enables the anticipation of potential problems and facilitates a rapid and effective response to emergencies.

From an economic perspective, the bibliometric analysis of digital actors reveals the significance of these technologies for business competitiveness and economic growth. Companies that adopt advanced technologies, such as machine learning and substitution models, can enhance their productivity and innovation capacity, thereby maintaining their competitiveness in an increasingly digitalized global market.

In a similar vein, these technologies have the potential to catalyze the emergence of novel business opportunities, thereby facilitating the development of markets and services that were previously unfeasible. This includes, but is not limited to, the development of predictive analysis platforms and maintenance services. The utilization of data-driven preventative services has the potential to enhance the sustainability of businesses. In the context of sustainability, the practical implications of this bibliometrics demonstrate the potential contributions of digital technologies to a more sustainable future, process optimization, and improvement in energy efficiency. This objective can be achieved through the use of simulation and models based on data, which have been shown to result in a substantial reduction of the carbon footprint for a variety of industries. In a similar vein, the capacity to monitor and administer resources in a more efficacious manner can assist in mitigating the effects of climate change and promoting more sustainable practices in production and consumption.

4.10. Limitations

It is imperative to acknowledge the limitations of this study when interpreting the findings. Firstly, the utilization of the Scopus and Web of Science databases suggests the possibility of bias in the results, as they may be disproportionately influenced by publications indexed on these platforms. This potential bias could result in the exclusion of relevant studies published in other repositories or in journals that are not indexed in these databases. This restriction may impose limitations on the generalizability of the findings and the comprehensive representation of the field of digital actors.

A further limitation is the utilization of tools such as Microsoft Excel® and VOSviewer® for data analysis. Despite the potency and pervasive utilization of these tools in bibliometric studies, the quality and precision of the results are contingent on the proper configuration and parameterization of analyses. Errors in data entry or parameter settings can lead to incorrect or incomplete interpretations. Furthermore, the identification and classification of growing and emerging keywords is influenced by temporal trends and data availability at the time of analysis. This could affect the reliability of these identifications in a dynamic and rapidly evolving context, such as that of digital actors.

4.11. Investigative Gaps

In the context of this bibliometric study on digital actors, Table 2 has been prepared to present in detail the main research gaps in this topic. These gaps represent key areas where the current literature reveals limitations and calls for future research to achieve a more complete and exhaustive understanding of the role and influence of digital actors in various contexts.

4.12. Research Agenda

The future research directions proposed in this section are directly derived from patterns identified in the bibliometric analysis, particularly from emerging and weakly connected nodes in the keyword co-occurrence network and from thematic areas located in peripheral quadrants of the frequency–validity map. For instance, terms associated with human-centered applications such as rehabilitation, therapeutic environments, and immersive interaction appear with relatively low frequency but high conceptual recency, suggesting underexplored intersections between performance-oriented digital twins and applied health domains. Similarly, the prominence of generic terms such as «machine learning» and «artificial intelligence» without strong linkage to performance metrics or adaptive behavioral modeling indicates a need for more specialized, context-driven algorithmic frameworks rather than general technical refinement. Therefore, the proposed agenda is not speculative but grounded in observable structural gaps, limited thematic integration, and emerging but insufficiently consolidated clusters identified in the bibliometric network analysis.

Although virtual reality is a relevant concept, it is only marginally integrated with performance-oriented digital twin frameworks. This finding indicates potential avenues for future research, namely the systematic integration of virtual reality with artificial intelligence and machine learning. Such integration could enhance the adaptive behavior, interaction quality, and performance evaluation of digital actors. The peripheral positioning of application-oriented terms related to human-centered domains indicates that areas such as rehabilitation, training, and therapeutic environments represent underexplored intersections rather than speculative extensions, particularly when analyzed through performance-driven digital actor frameworks.

Machine learning and deep learning manifest as prevailing yet vaguely delineated themes, exhibiting minimal explicit correlation with performance metrics and behavioral optimization of digital entities. This discrepancy underscores the necessity for more specialized, performance-aware learning methodologies that are meticulously designed to facilitate real-time adaptation, autonomy, and decision-making in digital characters and avatars. The bibliometric structure indicates that the advancement of these methods necessitates a transition from generic algorithmic improvements to context-specific learning strategies that are aligned with digital twin architectures.

The analysis also underscores the emergence of surrogate models and smart manufacturing as themes that, while developing, have yet to be fully consolidated. Their position in the thematic landscape indicates a need for research that strengthens the integration of surrogate modeling within digital twin–based digital actors to enable rapid performance prediction and optimization under real-time constraints. This is particularly salient in manufacturing and production environments, where performance evaluation and adaptability are pivotal yet have not been systematically addressed in the extant literature.

Finally, despite the robust post-2020 growth in scientific output, the bibliometric results indicate a paucity of longitudinal and comparative studies centered on production performance and system evolution. Based on these gaps, Figure 9 presents a research agenda that identifies potential avenues for future studies, including long-term performance assessment, interoperability between digital entities, the Internet of Things (IoT), data analytics, and the validation of digital twin–enabled decision-making across various domains. Addressing these gaps would contribute to a more mature and coherent research field, grounded in empirically validated and performance-oriented digital actor frameworks.

5. Conclusions

The bibliometric study on digital actors and digital twin technologies has provided a detailed and structured view of how this field of study has developed from its beginnings to the present. Research on digital actors shows exponential growth, evidenced by an 88.36 % increase in publications between 2020 and 2023. This increase not only reflects academic interest but also a growing recognition of the practical applicability of these technologies in various industries. The constant technological evolution and the need for more humanized interfaces in digital environments have placed this field at the forefront of scientific and technological innovation.

The analysis identified leaders in the field whose work has been fundamental to the development of knowledge on digital actors. These researchers, often working in collaboration with multiple institutions, not only contribute a significant volume of publications, but their research is widely cited and used as a basis for additional studies. This influence underscores the importance of mentoring and research collaboration to amplify both the academic and practical impact of this field.

The field has moved from initial explorations that integrated basic elements of performance and digitalization towards complexities such as machine learning and predictive models in the creation of digital actors. This transition not only demonstrates an improvement in technical and creative capacity, but also in the adaptation of these technologies to specific performance requirements and real needs, opening new avenues for explorations in augmented reality and interactive systems.

Regional differences in scientific output on digital actors highlight the role of economic resources and research policies in technological advancement. China and the United States lead this field, which may be linked to their substantial investments in technology and innovation, as well as their regulatory frameworks that support research and development in new technologies.

The interdisciplinary nature of research on digital actors reflects the convergence of computer science, animation, cognitive psychology, and ethics. Collaboration between experts from these disciplines is crucial to address emerging technical and ethical challenges, ensuring that technological developments are both innovative and responsible, and reflect a variety of humanistic and technological perspectives. Emerging keywords such as «artificial intelligence» and «deep learning» not only mark priority technologies but also expectations for future evolution in the field. These terms signal a growing interest in advanced automation and the ability of machines to learn and adapt to complex contexts of action, which could revolutionize how we interact with digital environments.

Despite notable progress, significant gaps remain in understanding the ethical, social, and privacy implications of digital actors. Furthermore, the adaptation of these technologies in non-Western contexts or in less technologically advanced industries presents a vast field for future research. These areas are of critical importance in ensuring that the development of digital actors is inclusive, ethical, and broadly beneficial.

These conclusions underscore the dynamics and importance of the field of digital players and digital twin technologies, offering a clear view of their advances and challenges. They also establish a framework for future research that can continue to explore and expand the boundaries of this area of study.

Conflict of Interest

Authors declare no conflicts of interest.

References

Alexander, O., Rogers, M., Lambeth, W., Chiang, J. Y., Ma, W. C., Wang, C. C., & Debevec, P. (2010). The digital emily project: Achieving a photorealistic digital actor. IEEE Computer Graphics and Applications, 30(4), 20–31. https://doi.org/10.1109/MCG.2010.65

Bastos Porsani, G., Lersundi, K., Sanchez-Ostiz Gutierrez, A., & Fernandez Bandera, C. (2021). Interoperability between building information modelling (BIM) and building energy model (BEM). Applied Sciences, 11(5), 2167. https://doi.org/10.3390/app11052167

Bordegoni, M., & Ferrise, F. (2023). Exploring the intersection of metaverse, digital twins, and artificial intelligence in training and maintenance. Journal of Computing and Information Science in Engineering, 23(6), 060806. https://doi.org/10.1115/1.4062455

Boussetta, M., Ababou, M., Faquir, S., & Rabiai, S. (2024). Artificial intelligence and digital infrastructure for sustainable industry and enhanced corporate performance: A bibliometric analysis of smart optimization technologies. En International Conference on Electrical Systems and Smart Technologies (pp. 93–102). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-86705-7_9

Chaudhary, R., Aslam, A. M., Bhardwaj, A., Kumar, N., & Buyya, R. (2024). Digital twins-enabled game theoretical models and techniques for metaverse connected and autonomous vehicles: A survey. SSRN. https://doi.org/10.2139/ssrn.4713492

Chua, P. C., Moon, S. K., Ng, Y. T., & Ng, H. Y. (2022). A surrogate model to predict production performance in digital twin-based smart manufacturing. Journal of Computing and Information Science in Engineering, 22(3), 031007. https://doi.org/10.1115/1.4053038

Fera, M., Greco, A., Caterino, M., Gerbino, S., Caputo, F., Macchiaroli, R., & D’Amato, E. (2019). Towards digital twin implementation for assessing production line performance and balancing. Sensors, 20(1), 97. https://doi.org/10.3390/s20010097

Guo, X., Liu, L., Huang, J., Wang, H., Du, X., Shi, J., & Wang, Y. (2023). A digital twin modeling method for array antenna assembly performance real-time analysis. The International Journal of Advanced Manufacturing Technology, 126(7), 3765–3781. https://doi.org/10.1007/s00170-023-11324-1

He, W., Mao, J., Song, K., Li, Z., Su, Y., Wang, Y., & Pan, X. (2023). Structural performance prediction based on the digital twin model: A battery bracket example. Reliability Engineering & System Safety, 229, 108874. https://doi.org/10.1016/j.ress.2022.108874

He, X., Lai, X., Yang, L., Zhang, F., Zhou, D., Song, X., & Sun, W. (2022). M-LFM: A multi-level fusion modeling method for shape-performance integrated digital twin of complex structure. Frontiers of Mechanical Engineering, 17(4), 52. https://doi.org/10.1007/s11465-022-0708-0

He, Y., Aliyu, A., Evans, M., & Luo, C. (2021). Health care cybersecurity challenges and solutions under the climate of COVID-19: Scoping review. Journal of Medical Internet Research, 23(4), e21747. https://doi.org/10.2196/21747

Heim, M. (1995). The design of virtual reality. Body & Society, 1(3-4), 65–77. https://doi.org/10.1177/1357034X95001003004

Hosamo, H. H., Nielsen, H. K., Kraniotis, D., Svennevig, P. R., & Svidt, K. (2023). Digital twin framework for automated fault source detection and prediction for comfort performance evaluation of existing non-residential Norwegian buildings. Energy and Buildings, 281, 112732. https://doi.org/10.1016/j.enbuild.2022.112732

Jamshidi, M. B., Sargolzaei, S., Foorginezhad, S., & Moztarzadeh, O. (2023). Metaverse and microorganism digital twins: A deep transfer learning approach. Applied Soft Computing, 147, 110798. https://doi.org/10.1016/j.asoc.2023.110798

Lai, J. Y., & Chang, L. T. (2021). Impacts of augmented reality apps on first graders' motivation and performance in English vocabulary learning. Sage Open, 11(4). https://doi.org/10.1177/21582440211047549

Li, L., Lei, B., & Mao, C. (2022). Digital twin in smart manufacturing. Journal of Industrial Information Integration, 26, 100289. https://doi.org/10.1016/j.jii.2021.100289

Li, Y. T., & Li, T. Y. (2008). A unified approach to planning versatile motions for an autonomous digital actor. Journal of Advanced Computational Intelligence and Intelligent Informatics, 12(3), 277–283. https://doi.org/10.20965/jaciii.2008.p0277

Liu, M., Fang, S., Dong, H., & Xu, C. (2020). Review of digital twin about concepts, technologies, and industrial applications. Journal of Manufacturing Systems, 58, 346–361. https://doi.org/10.1016/j.jmsy.2020.06.011

Longo, F., Nicoletti, L., & Padovano, A. (2019). Ubiquitous knowledge empowers the smart factory: The impacts of a service-oriented digital twin on enterprises’ performance. Annual Reviews in Control, 47, 221–236. https://doi.org/10.1016/j.arcontrol.2019.01.001

Lu, X., Shi, Z., Zhang, X., Cao, L., & Liang, S. (2025). Exploring the field of virtual avatar: A bibliometric and content analysis. International Journal of Human–Computer Interaction, 41(12), 7825–7841. https://doi.org/10.1080/10447318.2024.2400405

Massafra, A., Costantino, C., Predari, G., & Gulli, R. (2023). Building information modeling and building performance simulation-based decision support systems for improved built heritage operation. Sustainability, 15(14), 11240. https://doi.org/10.3390/su151411240

Mogensen, R. S., Rodriguez, I., Schou, C., Mortensen, S., & Sørensen, M. S. (2021). Evaluation of the impact of wireless communication in production via factory digital twins. Manufacturing Letters, 28, 1–5. https://doi.org/10.1016/j.mfglet.2021.01.006

Ofosu, R., Hosseinian-Far, A., & Sarwar, D. (2022). Digital twin technologies, architecture, and applications: A comprehensive systematic review and bibliometric analysis. En Blockchain and Other Emerging Technologies for Digital Business Strategies (pp. 105-142). Springer. https://doi.org/10.1007/978-3-030-98225-6_5

Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. International Journal of Surgery, 88, 105906. https://doi.org/10.1016/j.ijsu.2021.105906

Park, S. Y., Kim, S. J., Park, C. H., Kim, J., & Lee, D. Y. (2023). Data‐driven prediction models for forecasting multistep ahead profiles of mammalian cell culture toward bioprocess digital twins. Biotechnology and Bioengineering, 120(9), 2494–2508. https://doi.org/10.1002/bit.28405

Psarommatis, F. (2021). A generic methodology and a digital twin for zero defect manufacturing (ZDM) performance mapping towards design for ZDM. Journal of Manufacturing Systems, 59, 507–521. https://doi.org/10.1016/j.jmsy.2021.03.021

Qamsane, Y., Phillips, J. R., Savaglio, C., Warner, D., James, S. C., & Barton, K. (2022). Open process automation-and digital twin-based performance monitoring of a process manufacturing system. IEEE Access, 10, 60823–60835. https://doi.org/10.1109/ACCESS.2022.3179982

Qi, L., Xu, X., Wu, X., Ni, Q., Yuan, Y., & Zhang, X. (2023). Digital-twin-enabled 6G mobile network video streaming using mobile crowdsourcing. IEEE Journal on Selected Areas in Communications. https://doi.org/10.1109/JSAC.2023.3310077

Stary, C. (2023). Digital process twins as intelligent design technology for engineering metaverse/XR applications. Sustainability, 15(22), 16062. https://doi.org/10.3390/su152216062

Tan, Y. R., Hofmeister, M., Phua, S. Z., Brownbridge, G., Rustagi, K., Akroyd, J., & Kraft, M. (2024). Beyond connected digital twins—From GIS to the World Avatar. SSRN. https://doi.org/10.2139/ssrn.5072721

Vanderhorst, H. R., Heesom, D., & Yenneti, K. (2024). Technological advancements and the vision of a meta smart twin city. Technology in Society, 79, 102731. https://doi.org/10.1016/j.techsoc.2024.102731

Wallace, S. (2023). Digital twin and metaverse technologies, geospatial simulation and sensor fusion tools, and object perception and motion control algorithms in immersive hyper-connected virtual spaces. Linguistic and Philosophical Investigations, 22, 264–280. https://doi.org/10.22381/lpi22202316

Wang, J., Li, X., Wang, P., & Liu, Q. (2024). Bibliometric analysis of digital twin literature: A review of influencing factors and conceptual structure. Technology Analysis & Strategic Management, 36(1), 166-180. https://doi.org/10.1080/09537325.2022.2026320

Wang, S., Lai, X., He, X., Qiu, Y., & Song, X. (2022). Building a trustworthy product-level shape-performance integrated digital twin with multifidelity surrogate model. Journal of Mechanical Design, 144(3), 031703. https://doi.org/10.1115/1.4052390

Yang, L., Ni, S. T., Wang, Y., Yu, A., Lee, J. A., & Hui, P. (2024). Interoperability of the metaverse: A digital ecosystem perspective review. arXiv. https://doi.org/10.48550/arXiv.2403.05205

Yang, X., Ran, Y., Zhang, G., Wang, H., Mu, Z., & Zhi, S. (2022). A digital twin-driven hybrid approach for the prediction of performance degradation in transmission unit of CNC machine tool. Robotics and Computer-Integrated Manufacturing, 73, 102230. https://doi.org/10.1016/j.rcim.2021.102230

Yi, L., Glatt, M., Ehmsen, S., Duan, W., & Aurich, J. C. (2021). Process monitoring of economic and environmental performance of a material extrusion printer using an augmented reality-based digital twin. Additive Manufacturing, 48, 102388. https://doi.org/10.1016/j.addma.2021.102388

Zhou, Y., Xing, T., Song, Y., Li, Y., Zhu, X., Li, G., & Ding, S. (2021). Digital-twin-driven geometric optimization of centrifugal impeller with free-form blades for five-axis flank milling. Journal of Manufacturing Systems, 58, 22–35. https://doi.org/10.1016/j.jmsy.2020.06.019

Culbert, J. H., Hobert, A., Jahn, N. et al. (2025). Reference coverage analysis of OpenAlex compared to Web of Science and Scopus. Scientometrics, 130, 2475–2492. https://doi.org/10.1007/s11192-025-05293-3