Implementing network applications for 5G-enabled robots through the 5G-ERA platform

Abstract

Novel orchestration architectures for 5G networks have primarily focused on enhancing Quality of Service, yet have neglected to address Quality of Experience concerns. Consequently, these systems struggle with intent recognition and End-to-End interpretability, resulting in the possibility of suboptimal control policies being developed. The 5G-ERA project has proposed and demonstrated an AI-driven intent-based networking solution for autonomous robots to address this issue. Specifically, the proposed solution employs a workflow consisting of four tools - Action Sequence Generation, Network Intent Estimation, Resource Usage Forecasting, and OSM Control Policy Generation - to map an individual vertical action's intent to a global OSM control policy. The paper describes how the 5G-ERA platform enables the onboarding and control of 5G-enabled robots and how we demonstrate the platform’s capabilities through the project’s use cases.