SMART NEtworks and ServiceS for 2030 (SMARTNESS 2030)
Published October 28, 2025 | Version v1
Conference paper Open

Zero-Lag Smart Pipes for Smart Factories: AI-Driven Programmable Transport in Open RAN

Authors/Creators

Contributors

  • 1. ROR icon Universidade Federal de Goiás
  • 2. Federal University of São Carlos
  • 3. ROR icon Universidade do Vale do Rio dos Sinos

Description

This demonstration addresses a key open challenge in Open Radio Access Network (O-RAN) deployments: how to intelligently allocate Transport Network (TN) resources to ensure low-latency for mission-critical applications. The demo emulates a Smart Factory scenario where the time-sensitive control traffic of robotic arms competes with industrial camera broadband video streams. We propose an intelligent transport controller that combines network slicing, Adaptive Neuro-Fuzzy Inference System (ANFIS), and Federated Learning (FL) to dynamically prioritize traffic per slice. The architecture uses P4 switches for local queue monitoring and real-time resource scheduling. The integration with the O-RAN disaggregated stack is based on Open Air Interface (OAI). Experimental results demonstrate valuable load balancing and buffer occupation reduction in the O-RAN midhaul.

Files

2025327378-CNSM-PDFeXpress-Pass.pdf

Files (354.7 kB)

Name Size Download all
md5:fc9d348891897fc05f178486a4bd0f32
354.7 kB Preview Download

Additional details

Funding

Fundação de Amparo à Pesquisa do Estado de São Paulo
SMART NEtworks and ServiceS for 2030 (SMARTNESS) 21/00199-8