This deliverable establishes the internal governance and quality framework for the FREE6G-RadEdge project, defining shared rules for CTTC and the subcontracted partner to ensure high-quality outcomes and serving as a live document updated twice during the project. It outlines project roles, responsibilities, and contacts; describes internal communication tools, shared resources, and file-naming conventions; and defines quality assurance and control methods for key outputs such as deliverables and publications. It also classifies existing and project-generated research data, mapping them into a structured database, and summarizes dissemination activities alongside data management, archiving, preservation, and IPR procedures. In addition, it details compliance measures with GDPR and presents a consolidated risk analysis, including affected work packages, impact levels, and mitigation actions.

This deliverable additionally provides updated definitions of management roles and responsibilities, a detailed description of project resources including the shared collaboration space, and a substantially expanded Data Management Plan. The DMP now includes a structured inventory of all data types handled in the project, clear assignment of responsibilities for coordination, data ownership, access control, and technical support, and a comprehensive application of FAIR principles. It further details data security policies, organizational and technical controls, incident response, secure deletion, long-term preservation, and ethical and legal considerations. In addition, the deliverable formalizes quality control and quality assurance procedures not only for deliverables but also for publications, dissemination, and demonstration materials, and concludes with an updated risk analysis outlining identified risks and corresponding mitigation actions.

In addition to the elements already covered in earlier summaries, this deliverable reports significant updates related to project execution and societal impact. It documents delays across several technical activities, explains their technical causes, and formally justifies a six-month project extension, including updated planning for activities, deliverables, and milestones, while confirming that objectives, scope, and budget remain unchanged. It also provides a detailed assessment of the implementation and impact of the Gender Equality Plan, highlighting concrete achievements, monitoring indicators, and next steps aligned with Horizon Europe principles. Furthermore, it evaluates the Employment Plan, detailing measures taken to attract young talent, support STEM education, foster innovation, develop future skills, and stimulate direct and indirect employment. Finally, it updates the project risk analysis, identifying delays and architectural complexity as key risks, and confirms that mitigation actions and the approved extension have kept the project on track toward its objectives.

This deliverable defines the technical requirements for a Proof of Concept aimed at demonstrating the feasibility of a Cell-Free massive MIMO network implemented over an existing O-RAN architecture, including the measurements, KPIs, and target performance values needed to assess its suitability. It introduces the CF-mMIMO paradigm, the O-RAN architecture, and the project objectives, then details the network architectures, nodes, protocols, and interfaces for both CF-mMIMO and O-RAN systems. The document further specifies the PoC demonstrations at both the PHY layer and the DU–CU load-balancing level, describing the experimental setups, emulation tools, and evaluation metrics. Finally, it consolidates the conclusions that will serve as a foundation for subsequent specification and architecture deliverables.

This deliverable studies a cell-free uplink system where randomly located user equipments (UEs) transmit to distributed O-RUs connected via high-capacity fronthaul links to a central unit. Dynamic clustering allows each UE to be served by one or more O-RUs on the same resources, while alternative O-RU processing strategies, particularly Compute-and-Forward (CF) and Distributed Lossy Computation (DLC), are analyzed. The document provides a brief state-of-the-art overview and a theoretical analysis of the fundamental information-theoretic limits of CF and DLC strategies for scalability in cell-free uplink systems. Follow-up work is planned to explore CF/DLC-aware UE clustering using advanced machine learning algorithms.

This deliverable presents an initial report on hybrid MIMO fronthaul network design for cell-free massive MIMO architectures, targeting dense user deployments with uniform coverage, high data rates, and interference-free operation. It addresses the key challenge of implementing scalable and cost-effective fronthaul networks by proposing machine-learning–driven hybrid MIMO solutions as an alternative to fiber-based infrastructures, leveraging mmWave beamforming radios to achieve high-capacity and flexible fronthaul links. The document details ML-assisted hybrid analog–digital beamforming and beam-sharing mechanisms for point-to-multipoint and adaptive access point topologies, optimized array configurations and precoding strategies to minimize energy consumption under SINR constraints, and advanced AI/ML-based digital predistortion techniques to ensure robustness, reliability, and RF compliance of large antenna arrays. Overall, it summarizes the initial WP3 research outcomes on capacity, energy efficiency, and robustness, which will be further extended and validated in the final hybrid MIMO fronthaul design deliverable.

This deliverable proposes the integration of probabilistic forecasting techniques as a near–real-time xApp within the Open RAN architecture to enhance resource prediction and decision-making in future B5G and 6G networks. It reviews and compares probabilistic and single-point forecasting methods for estimating base station resource utilization, particularly PRB demand, demonstrating that probabilistic approaches provide more accurate, reliable, and informative predictions by capturing uncertainty. Through simulation-based evaluations, the work shows that DeepAR significantly outperforms traditional single-point methods such as LSTM and Seasonal-Naive baselines, as well as other probabilistic models. By embedding probabilistic forecasting into the O-RAN RAN Intelligent Controller, the deliverable highlights how improved CPU and PRB utilization prediction can support advanced RAN functions such as scheduling, beamforming, mobility management, and capacity planning, enabling more efficient, flexible, and vendor-neutral network operation.

This deliverable differs from the initial deliverable mainly in scope, depth, and maturity of the analysis. While the earlier deliverable provided an initial overview and preliminary theoretical insights into Compute-and-Forward and Distributed Lossy Computation for cell-free uplink systems, the current deliverable significantly extends this work with a more detailed and rigorous information-theoretic analysis, including achievable rate formulations. It places stronger emphasis on scalability issues, locality of CSI, and dynamic UE clustering under high-capacity fronthaul assumptions. In addition, this version more clearly positions CF and DLC as foundational mechanisms for future machine-learning–driven clustering and optimization, whereas the previous deliverable focused more on conceptual validation. Overall, the new deliverable consolidates and deepens the theoretical groundwork, preparing the path toward algorithmic and ML-based extensions in subsequent project phases.

This deliverable presents the final report on hybrid MIMO fronthaul network design, extending and consolidating the work initiated in D3.2 within the context of cell-free massive MIMO architectures for dense user deployments. It addresses the challenge of scalable and cost-effective fronthauling by proposing machine-learning–assisted mmWave hybrid beamforming solutions as alternatives to fiber-based infrastructures, targeting O-RAN deployments. The document details per-antenna power–constrained hybrid analog–digital beamforming and beam-sharing mechanisms for point-to-multipoint and adaptive access point topologies, evaluated through complexity analysis and numerical simulations. It further investigates energy-efficient array configurations, codebooks, and joint precoding strategies under SINR constraints, assessing scalability with respect to access points, antennas, and users, and exploring deployment within the RAN Intelligent Controller. Finally, the deliverable presents advanced AI/ML-assisted digital predistortion techniques, including decentralized and edge-assisted approaches, to ensure robustness, energy efficiency, and RF compliance of mmWave fronthaul links.

This deliverable investigates AI-enabled resource provisioning in cloud-native O-RAN environments, emphasizing the integration of probabilistic forecasting as a resource management rApp for future 6G networks. It analyzes cloud-native O-RAN architecture and rApp deployment options, highlighting how containerized solutions enable interoperable and flexible resource allocation. The work evaluates probabilistic forecasting techniques, particularly DeepAR and Transformer models, and compares them with deterministic approaches such as LSTM and simple feed-forward estimators. Through experimental analysis, it demonstrates that probabilistic models achieve lower prediction error while maintaining balanced CPU and memory usage, with DeepAR offering the best overall trade-off between accuracy and computational efficiency. The deliverable also examines the impact of data length on prediction accuracy and runtime and presents real-world case studies, concluding that probabilistic forecasting rApps significantly enhance intelligent, reliable, and scalable resource management in cloud-native O-RAN systems.

This deliverable documents the integration, testing, and deployment of a cell-free massive MIMO (CF-mMIMO) system over a disaggregated O-RAN architecture, demonstrating the practical feasibility of advanced 6G radio access concepts. Building on the CF-to-O-RAN mapping proposed in D3.4, the work addresses low-latency inter-DU communication and selective AP clustering for scalability. A flexible PHY proof-of-concept (PoC) testbed was developed, leveraging Keysight instrumentation including the MTRX multi-transceiver platform, PROPSIM RF channel emulator, and PathWave software suite, enabling emulation of multi-user, multi-AP scenarios with detailed channel modeling and beamforming. Performance validation shows improvements in SINR, block error rate, and error vector magnitude through joint transmission across APs. Practical challenges such as synchronization and receiver behavior at high SNR were identified, informing future refinements. The results confirm the technical viability of CF-mMIMO deployment in an O-RAN-compliant infrastructure. Future work will extend to dynamic scenarios with mobility, realistic digital twins, and distributed precoding techniques, establishing a foundation for scalable, robust next-generation wireless networks.

Presentation reporting on the successful completion of the demonstration “Demonstration of cell-free networking dense and ultra-dense hotspot areas” presenting the architecture, challenges, objectives, test setup with T&M equipment and SW tools, demonstration scenario and flow, propagation and mobility conditions, precoding, and SINR, EVM, and BLER results under imperfections. The document highlights the testbed as a valuable tool for future research in 6G wireless technologies, including dynamic scenarios with mobility, digital twin environments, and distributed precoding techniques for scalable and robust next-generation networks.

In this deliverable, the project documents all dissemination and communication efforts conducted during the first half of FREE6G-RadEdge. Activities targeted both technical and non-technical audiences to share project progress and results. A dedicated project website was created to provide accessible information and enable direct contact with the FREE6G teams, while social media channels, such as LinkedIn, were used to broaden outreach. Scientific dissemination included publications in journals, presentations at international conferences and workshops, invited talks, and supervision of Master’s and PhD theses. Other initiatives involved participation in industry events, contributions to associations, presentations in standardization groups, and organization of project-related events. The deliverable also highlights broader internationalization and impact-boosting activities aimed at raising project visibility. These coordinated efforts ensure that project outcomes are effectively communicated, supporting knowledge transfer, stakeholder engagement, and community building. Overall, D5.1 demonstrates a structured approach to both dissemination and communication, laying the foundation for continued visibility and impact throughout the project.

In this deliverable, the FREE6G-RadEdge project reports on dissemination and communication activities conducted during the second half of the project. Efforts focused on engaging both technical and non-technical audiences, ensuring broad visibility of the project’s objectives, outcomes, and societal impact. The project website was regularly updated, complemented by active outreach via professional social media channels to promote results and facilitate interaction with diverse stakeholders. Scientific dissemination included publications in international journals, conference presentations, invited talks, white and position papers, and academic contributions such as theses. Industry engagement was intensified through participation in major events, exhibitions, and workshops, strengthening liaisons with European players and promoting technology transfer. Contributions to standardization bodies and industry associations helped position the project’s innovations in the broader 6G ecosystem. Educational and public outreach initiatives targeted students and the general public, raising awareness of the societal benefits of 6G technologies. Overall, the deliverable highlights the strategic coordination of dissemination under WP5, emphasizing collaboration, internationalization, and impact maximization. The activities ensured that project outcomes reached relevant stakeholders, supporting both immediate and long-term visibility and uptake of FREE6G-RadEdge innovations.