Testbed Development: An Intelligent O-RAN based Cell-Free MIMO Network

TL;DR

This paper introduces a novel O-RAN-based testbed for cell-free MIMO networks that uses open source components and supports commercial 5G devices, enabling AI-driven network optimization.

Contribution

It presents a new CF-MIMO testbed leveraging O-RAN architecture with AI capabilities, capable of serving commercial 5G UEs and demonstrating network deployment and optimization.

Findings

Successful deployment of a CF-MIMO testbed with O-RAN architecture

Implementation of an AI-based antenna association xApp

Enhanced network performance through AI-driven optimization

Abstract

Cell-free multiple input multiple output (CF-MIMO) systems improve spectral and energy efficiencies using distributed access points (APs) to provide reliable service across an area equivalent to multiple conventional cells. This paper presents a novel design and implementation of a CF-MIMO network leveraging the open radio access network (O-RAN) architecture based testbed to enhance the performance of interference-prone user. The proposed prototype is developed based on open source software components and unlike many other prototypes, our testbed is able to serve commercial 5G user equipment (UE). The RAN intelligent controller (RIC) allows the cell-free (CF) network to access the embedded artificial intelligence and benefit from the network optimisation techniques that O-RAN brings. The testbed includes an intelligent antenna association xApp which determines the antenna group that serves each UE based on the live key performance measurements. The paper demonstrates the deployment and operation of the CF network and the xApp and discusses how the CF networks can benefit from the O-RAN architecture.