PoC 19 AI-enabled predictive closed loops: Difference between revisions
(Created page with "== PoC Team == * Nextworks * Orange Romania * CAFA Tech * Industrial System Institute == Main Contact == * Giacomo Bernini (g.bernini''_at_''nextworks.it) * Matteo Ravalli (m.ravalli''_at_''nextworks.it) == Contributors == * Catalin Brezeanu (catalin.brezeanu''_at_''orange.com) * Marius Iordache (marius.iordache''_at_''orange.com) * Märten Rannu (marten.rannu''_at_''cafatech.com) * Tanel Järvet (tanel.jarvet''_at_''cafatech.com) * Christos Verikoukis (cveri''_at_''isi...") |
No edit summary |
||
| Line 27: | Line 27: | ||
This PoC is WORKING on: | This PoC is WORKING on: | ||
* Topic 3 - [[Topic 3 - Intent-driven Closed-Loop automation|Intent-driven Closed-Loop automation]] | * Topic 3 - [[Topic 3 - Intent-driven Closed-Loop automation|Intent-driven Closed-Loop automation]] | ||
* Topic 2 - [[ | * Topic 2 - [[Topic2 - Automation in Multi-Stakeholder Ecosystems|Automation in Multi-Stakeholder Ecosystems]] | ||
== PoC Demo == | == PoC Demo == | ||
Latest revision as of 16:27, 20 April 2026
PoC Team
- Nextworks
- Orange Romania
- CAFA Tech
- Industrial System Institute
Main Contact
- Giacomo Bernini (g.bernini_at_nextworks.it)
- Matteo Ravalli (m.ravalli_at_nextworks.it)
Contributors
- Catalin Brezeanu (catalin.brezeanu_at_orange.com)
- Marius Iordache (marius.iordache_at_orange.com)
- Märten Rannu (marten.rannu_at_cafatech.com)
- Tanel Järvet (tanel.jarvet_at_cafatech.com)
- Christos Verikoukis (cveri_at_isi.gr)
Abstract
This PoC will showcase a zero-touch network and service management solution, capable to deploy virtualized AI-enabled closed loop functions in support of prediction of network slice performance degradation in collaborative robots application scenarios. The PoC targets the AI-enabled and automated operation of vertical services and network slices through zero-touch closed loops, with deployment of services for prioritized traffic (URLLC) in support of collaborative robot operations. The goal is to validate the zero-touch deployment and operation of virtualized closed loop functions, integrated with an MLOps framework for training and deployment of ML models, that make use of ML-enabled analytics (for prediction of network slice performance) and ML-enabled decision (for network slice congestion identification and remediation) to guarantee ultra low latency communications for collaborative robots.
This PoC is developed using the network and service management solutions, including an MLOps platform for lifecycle management of AI/ML models, developed in the European Commission (EC) funded project, under the Horizon Europe programme, ADROIT6G [1]. The implementation of these solutions is aligned with the ZSM standards. More specifically, the design architecture of the use case follows the automation principles of ZSM 003, while the AI-driven closed loop based automation, which integrates a closed loop governance and an MLOps platform, is aligned with the ZSM 009-1, ZSM 009-2 and ZSM 009-3.
PoC Proposal
Full PoC proposal on ETSI portal as ZSM(26)000050r1 - Download pdf
PoC Topics
This PoC is WORKING on:
- Topic 3 - Intent-driven Closed-Loop automation
- Topic 2 - Automation in Multi-Stakeholder Ecosystems
PoC Demo
- demo #1, June 2026, Showcasing PoC results at 2026 EuCNC & 6G Summit conference.
- demo #2, July 2026, Showcasing Final PoC results at 2026 IEEE MeditCom conference.
PoC Report(s)
- PoC #19 Final report expected in July 2026