Collaborative edge-cLoud continuum and Embedded AI for a Visionary industry of thE futuRe

Informazioni relative al progetto

CLEVER

ID dell’accordo di sovvenzione: 101097560

DOI

10.3030/101097560

Progetto chiuso

Data della firma CE 29 Novembre 2022

Data di avvio 1 Gennaio 2023

Data di completamento 31 Dicembre 2025

Finanziato da

• Digital, Industry and Space

Costo totale

€ 14 974 655,00

Contributo UE

€ 4 497 177,77

Coordinato da

SCUOLA SUPERIORE DI STUDI UNIVERSITARI E DI PERFEZIONAMENTO S ANNA
Italy

Periodic Reporting for period 3 - CLEVER (Collaborative edge-cLoud continuum and Embedded AI for a Visionary industry of thE futuRe)

Periodo di rendicontazione: 2025-01-01 al 2025-12-31

Sintesi del contesto e degli obiettivi generali del progetto

The CLEVER project is focused on developing processing solutions for AI at the edge, addressing the design stack and middleware. It aims to devise a platform that suits specific applications by leveraging technical advancements. The technical objectives include defining architectural frameworks, functional requirements, and Key Performance Indicators (KPIs) for the CLEVER solution and its use cases. Additionally, the project aims to specify AI solutions and enabling technologies for Digital Twin, Smart Agriculture, and Augmented Reality for Shopping. The impact of the project is expected to be demonstrated through communication and dissemination activities, with the plan for exploitation and dissemination of results outlined to maximize the project's impact.

The project's overall objectives include setting up the environment for project management, communication, and dissemination, as well as better specifying the CLEVER architecture and the three considered use cases and the necessary hardware and software tools to support them. The expected impacts of the project revolve around advancing knowledge and results, implementing them into advanced curricula, and disseminating project results through publications in international peer-reviewed journals. Additionally, the project aims to communicate information with respective supply chains and monitor its impact to ensure commercial exploitation evolves throughout the project.

Lavoro eseguito dall’inizio del progetto fino alla fine del periodo coperto dalla relazione e principali risultati finora ottenuti

During the project, substantial technical and scientific work was carried out to design, implement, integrate, and validate an advanced cloud–edge continuum for AI‑enabled applications. The performed activities covered the development of secure onboarding and data trust mechanisms based on distributed ledger technologies and remote attestation, AI‑driven orchestration supported by distributed knowledge graphs, and multi‑objective resource allocation algorithms targeting both performance and energy efficiency. In parallel, novel hardware accelerators—including AI system‑on‑chip platforms, neuromorphic and optical computing engines, and RDMA‑enabled data processing units—were developed and integrated with the middleware and application layers. These efforts resulted in a fully integrated and experimentally validated solution, demonstrated across multiple industrial use cases, which achieved low‑latency execution, efficient AI workload offloading, secure and trustworthy data handling, and energy savings of up to approximately 30%. The final system reached Technology Readiness Level 6–7, confirming the maturity, robustness, and industrial relevance of the achieved technical results.

Progressi oltre lo stato dell’arte e potenziale impatto previsto (incluso l’impatto socioeconomico e le implicazioni sociali più ampie del progetto fino ad ora)

The project delivered validated technical results for an AI‑enabled cloud–edge continuum, combining secure middleware, AI‑driven orchestration, and heterogeneous hardware acceleration. The main outcomes include trustworthy device onboarding and data confidence mechanisms, distributed knowledge‑graph‑based orchestration, and energy‑aware resource management, together with the integration of AI system‑on‑chip platforms, neuromorphic and optical accelerators, and RDMA‑enabled data processing units. These results were integrated into a unified architecture and validated in industrially relevant use cases, demonstrating low latency, efficient AI workload offloading, secure data handling, and energy savings of up to around 30%, with the overall system reaching TRL 6–7.
These results have strong potential impact by enabling scalable, secure, and energy‑efficient deployment of AI services across heterogeneous edge and cloud infrastructures. They support European objectives in sustainability, digital sovereignty, and industrial competitiveness, and are directly applicable to latency‑sensitive and data‑intensive sectors such as manufacturing, agriculture, and immersive digital services, while remaining transferable to a broader range of industrial domains.
Further uptake and long‑term success require additional large‑scale demonstrations and pilot deployments, continued applied research to progress selected components towards full industrial maturity, and clear exploitation and IPR strategies. Access to market‑facing ecosystems, early adopters, and appropriate financing mechanisms will be essential to support commercialisation and internationalisation, alongside continued alignment with evolving regulatory requirements and standardisation frameworks to ensure interoperability and wide adoption.