The FREE6G_RadEdge team member Adriano Pastore participated in the EuCNC’25 exhibition demonstration entitled “Neural Constellation Shaping for Over-the-Air Communication”.

 Non-Orthogonal Multiple Access (NOMA) enables simultaneous transmissions over shared frequency resources, improving spectrum efficiency but requiring advanced interference cancellation. Conventional power-domain NOMA with standard modulations underutilizes the signal space and limits performance. End-to-end deep learning, particularly autoencoder-based methods, enables joint modulation and coding optimization and data-driven constellation shaping, improving throughput, spectrum efficiency, and error rates for advanced 6G applications such as federated learning and distributed sensing.

The researcher Adriano Pastore participated in the IEEE International Conference on Machine Learning for Communication and Networking (ICMLCN) 2025 in Barcelona showcasing part of the work conducted in the FREE6G-RadEdge project. 

He delivered the demonstration "Autoencoder-Based Dynamic Coded Modulation for 6G NOMA Uplink" showcasing a NOMA uplink implemented with software-defined radios that demonstrates dynamic link adaptation using an end-to-end learned, coded modulation system. A model-free approach to training using over-the-air samples helps illustrate the superiority of neural implementations of mappers and demappers that can factor in the myriad hardware impairments encountered in practical communication systems. 

The FREE6G-RegEdge team, on this occasion represented by the partner Nearby Computing, presented the work "Joint UPF and Application Placement in Multi-Slice Edge Networks: A Reinforcement Learning Strategy" at the 2025 IEEE Wireless Communications and Networking Conference (WCNC).

What is it about? The virtualization of 5G/6G networks enables cloud-native deployment of network and application functions, where the placement of user plane functions and services strongly affects end-to-end latency. In multi-slice edge environments, this placement problem becomes increasingly complex due to heterogeneous user requirements, multiple cost factors, and limited edge resources. The study addresses joint UPF and application placement by considering service migration, traffic forwarding, server activation, and processing costs. To handle this complexity, a reinforcement learning–based approach, JUAP-RL, is proposed to jointly decide and coordinate UPF and application deployment locations.

This year, we had the privilege of having Keysight Technologies showcase a video highlighting their contribution to the FREE6G Radio Edge project at the main Keysight booth during MWC2025. Building on this video, Carles Navarro Manchón presented the testbed developed within the framework of the project, leveraging Keysight’s advanced test and measurement (T&M) products, software solutions, and innovative adaptations to meet the requirements of cell-free scenarios, as well as the main results achieved by the project.

Following the presentation, Nikolaos Bartzoudis, PI of the FREE6G-RadEdge project, interviewed Michael Dieudonne, R&D Manager at Keysight, and Carles Navarro on the project’s exploitation strategy and the opportunities enabled by its outcomes. The Keysight team highlighted how the project allowed them to gain expertise in a new technology that is already being integrated into their product portfolio, as well as to attract talent, strengthen their presence in Spain, and position the country as a hub for this type of advanced developments.

Special thanks to the Keysight team for their availability during the project presentation and the interview.

Nikolaos Bartzoudis, PI of FREE6G-Security, had the opportunity to participate in the panel session “Innovating for the Future of AI and 6G Networks,” organized by NTT DATA, a partner of the project, at its global stand for the MWC25 event. The session aimed to explore cutting-edge developments in EdgeCloud, the transformative potential of blockchain, intelligence in autonomous networks with O-RAN and SMO, and the role of AI in enhancing all these technologies. Alejandro Fernández, Telecom Director at NTT DATA (acting as moderator), Diego Rueda, Telecom Expert Consultant at NTT DATA, Shujaur Mufti, Director of Global Telco Ecosystem Architecture, and Nikolaos Bartzoudis engaged in insightful discussions and shared forward-looking perspectives on the key enablers and expected outcomes of these innovations for the future connected society.

Following the panel session, Alejandro Fernández presented the main outcomes of the FREE6G-Security project. Nikolaos Bartzoudis also interviewed him on the project’s exploitation strategy and short-term initiatives emerging from its results. We thank NTT DATA for the time and visibility given to the FREE6G-Security project.

At MWC25 we had the opportunity to discuss the participation of Nearby Computing and Cellnex in the FREE6G Regional Edge project. Angelos Antonopoulos, Research and Innovation Director of Nearby Computing, first introduced the main solutions on Elastic MEC Platform and Zero-touch Slice orchestration, the architecture and results (achieved functionalities and benefits of the proposed AI/ML orchestration and slicing algorithms) delivered by project.

After the project presentation, Nikolaos Bartzoudis, PI of FREE6G-RegEdge interviewed Angelos Antonopoulos, and Carmen Vicente, Project Manager and IoT Engineering, and Javier Santaella Sánchez, Innovation Project Manager, both at Cellnex on the main take aways of the project, how could it reinforce the Spanish 5G company ecosystem, planned exploitation of results and how the relationship with R&I organizations like CTTC could be further exploited.

Special thanks to them all for their availability during the project presentation and the interview.

Nikolaos Bartzoudis, PI of the FREE6G projects, showcased a live demo at CTTC booth based on real-time computer vision-assisted reconfiguration of SoC accelerated functions.

SoC telemetry enables KPI-driven reconfiguration to optimize energy and latency, while efficient function placement and offloading improves resource utilization, supports multi-tenancy at the edge, and contributes to sustainable computing.

This work is part of the technologies being explored at the FREE6G-RegEdge project and gave visibility of the project to telecom companies and industry and public-private partnership associations representatives like the SNS JU, and both national and EC authorities.

Once again, the research team of the FREE6G project participated in Catalan Science Week 2024, this time hosting a series of exciting talks and demonstrations on 5G and mmWave communications.

Held at CTTC’s Wireless Communication Lab, David López led the presentations, providing in-depth insights into the cutting-edge research behind FREE6G. The talks were delivered to high-school students (ESO 4th course) from Institut Escola Gavà Mar and students enrolled in the IT formative cycle at Escola Pàlcam.

An incredible opportunity to inspire the next generation of engineers and spark interest in the rapidly evolving field of telecommunications!

The projects FREE6G-RadEdge and FREE6G-RegEdge team up and present the work "Minimizing Power Consumption under SINR Constraints for Cell-Free Massive MIMO in O-RAN", led by the project researchers Miguel Ángel Vázquez and Luis Blanco, at the 2024 20th International Conference on Network and Service Management (CNSM).

This work addresses energy consumption minimization in Open RAN cell-free massive MIMO systems while guaranteeing minimum SINR requirements for all users. Multiple multi-antenna access points jointly serve users over shared resources, and the objective is to design precoding vectors that reduce total power consumption, accounting for both fixed access point activation costs and variable transmit power. The resulting formulation is a strongly non-convex, binary-constrained quadratic optimization problem. To solve it, the paper proposes a novel penalized convex–concave procedure that jointly optimizes precoding and access point selection. The solution is suitable for implementation as an xApp in the near–real-time RAN Intelligent Controller, and numerical results demonstrate significant energy efficiency gains compared to baseline approaches.

The FREE6G-Security project was present at the Reunión Española de Criptología y Seguridad de la Información (RECSI) 2024 in León (Spain), represented by partner Universitat Oberta de Catalunya, with the work "Unveiling Privacy Risks: Property Inference
Attacks in Federated Learning Environments"

Federated Learning enables privacy-preserving model training by keeping data localized on devices, but it remains vulnerable to certain privacy attacks. This FREE6G-Security work examines property inference attacks, where adversaries infer hidden attributes of the training data that are unrelated to the main learning task. Using TensorFlow Federated, the study reproduces and extends prior results, confirming the robustness and credibility of such attacks. The analysis shows that attacks are most effective during early training rounds due to large initial model updates, and that rarer properties are easier to infer because they create stronger weight contrasts. The findings demonstrate that significant privacy risks exist in FL systems even without adversarial manipulation, underscoring the need for stronger protection mechanisms.

A first effort towards introducing the concept of the FREE6G-RegEdge micro-orchestration of RAN functions hosted in FPGA SoC devices was presented last week in the IEEE 6GNet'24 conference by Nikolaos Bartzoudis, PI of FREE6G.

Micro-orchestration stands for function scaling, migration (within the SoC device), replacement or reconfiguration. The paper presents a FPGA SoC function micro-orchestration driven by computer vision events (i.e., face detection + face counting) and is currently extended to account for other multimodal data (e.g., RAN traffic, FPGA SoC deep telemetry data, on-chip or ambient sensors) that will be digested by an AI/ML model in the form of an rApp, xApp or dApp.

This is a joint work of Centre Tecnològic de Telecomunicacions de Catalunya (CTTC) and Nearby Computing.

The FREE6G project was present at the EuCNC conference held in Antwerp, Belgium, with the paper “Enhancing Cloud-Native Resource Allocation with Probabilistic Forecasting Techniques in O-RAN”. This work has been led by Luis Blanco in the frame of the FREE6G-RadEdge and FREE6G-RegEdge projects. The cloud-native aspects of O-RAN together with the radio App (rApp) deployment options are examined. The integration of probabilistic forecasting techniques as a rApp in O-RAN is also emphasized, along with case studies of real-world applications. 

In addition, David López-Bueno imparted the invited talk "Sustainable AI for Trustworthiness and Energy-Proportional Radio Micro Orchestration", at the workshop "Sixtainability: Paving the Way for a Sustainable 6G" on behalf of the SNS VERGE project, one of the key EU projects spun out from concepts being researched at FREE6G like the radio and distributed units computing and radio resource micro-orchestration leveraging RAN intelligence, Edge sensors, and Edge to radio unit observability.

Researchers from the FREE6G team presented the project at MWC 2024 in Barcelona, showcased at the booth of the Centre Tecnològic de Telecomunicacions de Catalunya (CTTC).

This UNICO I+D 5G project is funded by the Ministerio de Asuntos Económicos y Transformación Digital through the Spain Recovery and Resilience Plan – NextGenerationEU mechanism.

The project was also highlighted within the Teleco Renta initiative, which aims to promote and encourage vocations and careers in telecommunication engineering.

Advancing 6G research while inspiring the next generation of telecom engineers!

The Nearby Computing team presented the FREE6G-RegEdge project during MWC 2024 in Barcelona.

The FREE6G-RegEdge team showcased the latest advances in elastic edge computing for 6G, including cloud-native MEC deployment, autonomous AI-driven orchestration, and zero-touch network slice reconfiguration. Together with Cellnex, these innovations will be demonstrated live at the Castellolí circuit testbed, highlighting near-zero-latency service delivery, distributed machine-learning-based control, and end-to-end integration of edge, core, and 5G/6G network functions validated under real traffic conditions.

The FREE6G team is happy to have participated in the Catalan Science Week '23. We have imparted a talk uncovering several aspects on capacity, interference, energy efficiency and sustainability, in radio equipment and 6G mobile communication networks.

The talk was given to students of the formative cycle in IT systems, networks and cybersecurity from Escola Pàlcam who visited the Centre Tecnològic de Telecomunicacions de Catalunya (CTTC). We appreciate very much their genuine interest and insightful questions.

The CTTC Researchers David López-Bueno and Antonio Román Villaroel have presented FREE6G at the Smart City Expo World Congress.

Fostering the future 6G-Enabled Smart Cities and other demanding verticals, the project aims at demonstrating scalable, distributed, machine learning-based  cell-free networks to support massive AP deployments, deploy an elastic edge computing paradigms, and deliver self-driven infrastructures leveraging policy-driven security, privacy and trust to support multiple tenants.

David López-Bueno, researcher of FREE6G-RadEdge, successfully defended the PhD Thesis entitled "Machine Learning Techniques for Adaptive Polynomial and Neural Network Digital Predistorters".

The thesis focuses on AI/ML techniques to reduce the complexity and training time of 5G and beyond-5G digital predistorters (DPDs). DPDs are essential digital front-end signal processing blocks used to improve the energy efficiency and sustainability of radio units. These aspects have been investigated within the framework of the FREE6G-RadEdge project, particularly in the exploration of mmWave fronthauling, but may have broader impact also on base station, user equipment and wireless backhaul radios.

The FREE6G team participated in a hybrid mini workshop at the University of Novi Sad in Serbia organized by Professor Dejan Vukobratovic. The workshop entitled “University of Novi Sad and CTTC mini colloquium on B5G R&D topics and future synergies”, was supported by the H2020 project INCOMING (CSA).

The goals and on-going R&D activities of FREE6G were presented to different research groups at University of Novi Sad to disseminate the project, raise interest and awareness and explore collaboration strategies. Members of the FREE6G-RadEdge and FREE6G-RegEdge teams imparted the two focused talks "Micro-orchestration of RAN VNFs: context and KPI-driven zero-touch adaptation of DU-RU functions" and "Machine learning techniques for adaptive polynomial and neural network digital predistorters", respectively.

FREE6G-Security presented the work "On the Decentralization of Blockchain-enabled Asynchronous Federated Learning", led by the project researcher Paolo Dini, at the 2023 IEEE 9th International Conference on Network Softwarization (NetSoft) conference.  

Why this work? Federated learning is expected to support collaborative intelligence at the edge, but its reliance on centralized orchestration raises concerns about security, privacy, and scalability. Blockchain has been proposed to decentralize federated learning, improving robustness and trust, leading to the FLchain paradigm. However, blockchain decentralization introduces challenges such as ledger forks, model inconsistencies, and stale updates caused by asynchronous operation. These effects can result in devices training on different or outdated global models, impacting learning performance. This work analyzes how blockchain-enabled decentralization influences the age of information and the accuracy of federated learning. To support this analysis, a simulation framework is developed to realistically capture the decentralized and asynchronous behavior of FLchain systems.

In the frame of an institutional visit at CTTC from the Generalitat de Catalunya Research and Universities Department, led by the Honorable Minister of Research and Universities Gemma Geis, together with its General Research Directorate, the FREE6G project was presented as a key example of cutting-edge telecom research.

During the visit, David López, researcher of the FREE6G project, delivered an informative talk on Beyond 5G research, highlighting how the project contributes to enhancing capacity, ubiquity, and energy efficiency in future radio access technologies.

The CTTC Workshop is a biannual event that aims at gathering all CTTC researchers, so that they can disseminate their technical work and share their ideas with other CTTC colleagues. Several vibrant poster sessions are organized in this respect. The Workshop hosts two renown invited speakers, along with stakeholders of the local hi-tech ecosystem.

Members of CTTC’s FREE6g team participated in the poster session presenting the project activities to boost cross-fertilization with other CTTC researchers and UNICO I+D 5G 2021 projects.