EXplainable Automation for Continuum 6G infrastructures

Project Information

EXACT-6G

Grant agreement ID: 101120297

DOI

10.3030/101120297

EC signature date 27 June 2023

Start date 1 December 2023

End date 31 May 2028

Funded under

• Marie Skłodowska-Curie Actions (MSCA)

Total cost

No data

EU contribution

€ 2 657 167,20

Coordinated by

ATHINA-EREVNITIKO KENTRO KAINOTOMIAS STIS TECHNOLOGIES TIS PLIROFORIAS, TON EPIKOINONION KAI TIS GNOSIS
Greece

Periodic Reporting for period 1 - EXACT-6G (EXplainable Automation for Continuum 6G infrastructures)

Reporting period: 2023-12-01 to 2025-11-30

Summary of the context and overall objectives of the project

EXACT-6G is an MSCA Doctoral Network which proposes a novel and agile 6G system architecture based on enabling technologies that combine RIS with a Compute Continuum infrastructure and Distributed, Intelligent orchestration. The novel paradigm seamlessly integrates Cloud, Edge and Far Edge nodes under a self-synthesized and self-managed PaaS system. It introduces new concepts, such as Pervasive Orchestration, and designs new Explainable AI models for scalable, autonomous control of the infrastructure, and allows microservices to “follow” users, being tracked by Joint Communication and Sensing mechanisms.
The project brings together a consortium from 4 EU countries which consists of three academic and three industrial beneficiaries, as well as three associate partners that will award the PhD titles.
Briefly, EXACT-6G will train, through their individual research projects, ten Doctoral Candidates who will deal with the challenges of i) Integrated communications and sensing in RIS ii) Orchestration, Network management and explainability and iii) Cloud continuum.
The three main technical objectives of the EXACT-6G research program are:
1. Develop data-driven optimization and localization algorithms for RIS to improve 6G KPIs for communications and sensing as well as the energy efficiency of the whole network operation. The proposed innovations and data-driven algorithms aim to help reduce the computational complexity typically required for RIS optimization and beam configuration, while producing realistic radio propagation models with nonideal hardware, increasing Spectral Efficiency (SE).
2. Design a Distributed and Intelligent Management /Orchestration plane (DIMO) that automates and disaggregates service and network management across administrative and technological domains, leveraging Explainable AI as a key enabler. Distributed Intelligent orchestration is envisioned to facilitate (micro-)services to be deployed over the CEC either using the resources provided under the governance of a single DIMO or over a federated CEC managed by different DIMO.
3. Develop a self-synthesizing continuum solution, supporting AI-driven, dynamic resource pool management and unified service deployment and orchestration from the cloud to the far edge.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

During the first period of the project,the following main activities were performed relatively to:

• WP4-Training:
- EXACT-6G Scientific schools 1and 2 were organized online in M16 while Schools 3 and 5 took place in Eurecom in M22,
- DCs were enrolled to PhD programs and started their individual research projects,
- organization of frequent meetings between the DCs and their supervisors.

• WP5- Management:
- the EXACT-6G website was published and the LinkedIn official account was launched and regularly fed with original posts,
- the recruitment process was formally initiated with the publication of the 10 DC positions in Euraxess and concluded in M22 with the recruitment of the final DC,
- the project's data management plan has been drafted and submitted,
- all relevant deliverables were timely submitted.

• WP6-Dissemination, Communication, and Standardization
- the first brochure, newsletter and dissemination video of the project were released,
- the 1st workshop was organized in the framework of the IEEE NFV-SDN 2025 Conference in Athens, Greece, 10-12 November 2025,

At the technical level, during the first reporting period the DCs worked on the development of the beyond state-of-the-art approach that they will later follow in their individual research projects.This preparatory work was described in Deliverables D1.1 D2.1 and D3.1 while it is also presented in papers published in journals and conference proceedings.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

The beyond state-of-the-art research contributions of EXACT-6G were obtained through the 3 technical WPs (WP1, WP2, WP3).

WP1, Data-Driven optimization, control and localization for RIS in 6G networks. The beyond state-of-the-art contributions for the first reporting period are listed below:
- Study of transmit beamforming and hybrid reconfigurable intelligent surface (HRIS) reflection and amplification coefficients optimization for maximizing spectral efficiency.
- Direction of arrival (DoA) estimation, which is a core problem in many beamforming problems including integrated sensing and communications (ISAC), has been solved via deep unfolding-based machine learning solution.
- RIS-assisted wireless systems have been considered under realistic propagation and interference conditions.

WP2, Distributed and Explainable Management and Orchestration. The beyond state-of-the-art contributions for the first reporting period are listed below:
- Implementation of a DIMO infrastructure for CEC resources discovery, prediction and composition.
- Development of an Inter-DIMO intent and policy propagation and monitoring approach, through the development of a Service Level Agreement (SLA) intent translation and intent propagation mechanisms within DIMO-based solutions for 6G wireless networks.
- Proposing a solution to the problem of handover decision-making in dense MIMO wireless networks, by formulating the handover problem as an RL task.
- Targeting full network orchestration automation through a Zero-touch Service Management (ZSM) approach, using an explainable AI (XAI) -driven ZSM closed-control loop scheme.

WP3, Self-managed Compute Continuum infrastructure for Metaverse services. The beyond state-of-the-art contributions for the first reporting period are listed below:
- Work has begun on the intent-driven PaaS system, with advances on the design of mechanisms for resource allocation and on the improvement of their scalability, robustness, and responsiveness.
- The first notions of metaverse applications for the compute continuum have been developed with the design and modelling a hierarchical 6G edge abstraction with heterogeneous availability of resources.