Semantic Interoperability for Autonomous Data Ecosystems A Systematic Review of Technologies, Challenges, and Future Directions

<title>Abstract</title> <p> <bold>Background</bold> Autonomous industrial and cyber-physical environments increasingly rely on heterogeneous machines, sensors, software services, and data platforms that exchange information with minimal human intervention. However, communication mechanisms alone enable only data transfer and do not guarantee that systems interpret exchanged information consistently. Semantic interoperability addresses this challenge by enabling machine-interpretable representation of data, entities, relationships, and contextual information. <bold>Objective</bold> This review examines semantic representation approaches, enabling technologies, context representation mechanisms, machine-to-machine communication, technical challenges, and unresolved research gaps associated with semantic interoperability in autonomous data ecosystems. <bold>Method</bold> A PRISMA 2020-based systematic literature review was conducted around five research questions covering semantic representation, interoperability technologies, contextual information, technical barriers, and future research directions. The review synthesizes peer-reviewed studies addressing semantic Web technologies, ontologies, knowledge graphs, Industry 4.0, Industrial IoT, cyber-physical systems, digital twins, OPC UA, and related communication approaches. <bold>Results</bold> The evidence highlights the widespread adoption of ontologies and semantic models for representing heterogeneous industrial information. Knowledge graphs increasingly extend these approaches by capturing relationships among assets, processes, and information resources. Digital twins and Industry 4.0 environments provide important application contexts for semantic integration, while OPC UA contributes both communication capabilities and information modelling support. However, significant challenges remain, including semantic heterogeneity, ontology alignment, model evolution, scalability, contextual inconsistency, standardization, and limited industrial-scale validation. <bold>Conclusions</bold> Semantic representation has become an essential mechanism for enabling interoperability among heterogeneous autonomous systems. Nevertheless, current evidence remains fragmented across domains, standards, and deployment environments. Key unresolved challenges include cross-domain semantic alignment, context interoperability, scalable knowledge representation, and standardized evaluation of semantic interoperability performance. </p>