In the context of recent developments in logistics, the efficient and flexible deployment of autonomous robotic systems remains a significant challenge, particularly in semi-structured, flexible environments, typically encountered in small residences, warehouses, or medium-sized industrial facilities, where there is minimal potential for infrastructural and/or procedural enhancements to facilitate robotic automation. Such environments pose a significant problem to automated solutions, since the environment is flexible and partially unknown, cluttered and human-center. This paper presents a modular architecture for autonomous robotic logistics designed to enhance operational efficiency through adaptive control, real-time decision-making, and IoT integration. The proposed system employs a hierarchical architecture that separates high-level task planning from lowlevel motion control, facilitating scalability and simplifying task management. The architecture supports multiple autonomous robots capable of dynamic task allocation, path planning, and predictive control to improve reliability and minimise task execution time. This work contributes to the development of a novel class of robotic logistics systems that are capable of operating in semi-structured environments. These systems are distinguished by their ability to combine the advantages of traditional large-scale automation solutions with the flexibility and cost-effectiveness of small-scale robotic systems.

Industry 5.0 improves human-machine collaboration by integrating Artificial Intelligence (AI) and Extended Reality (XR) into industrial environments. This paper presents a structured Reference Architecture for a system that addresses industrial environments within the Industry 5.0 paradigm, iteratively refined based on evolving business and technical requirements. Designed for scalability, interoperability, and flexibility, it enables seamless interaction between XR applications, AI-powered decision support tools, and industrial systems. The architecture incorporates an enterprise architecture perspective that addresses industry-driven use cases and user scenarios from six pilot applications. Key components, including XR applications, AI recognition models, digital twins (DTs), and an orchestration hub, are outlined. The business layer is built on user stories that reflect real-world industrial needs from various sectors, and future directions for extending the architecture within Industry 5.0 ecosystems are discussed.