Table of Contents
Fetching ...

Building a real-time physical layer labeled data logging facility for 6G research

Franco Minucci, Raquel Marina Noguera Oishi, Haoqiu Xiong, Dieter Verbruggen, Cel Thys, Rizqi Hersyandika, Robbert Beerten, Achiel Colpaert, Vida Ranjbar, Sofie Pollin

TL;DR

The architecture and vision of designing and implementing a new test infrastructure for 6G physical layer research at KU Leuven and how the Testbed can be used for research on joint communication and sensing, over-the-air synchronization, distributed processing, and AI in the loop are described.

Abstract

This work describes the architecture and vision of designing and implementing a new test infrastructure for 6G physical layer research at KU Leuven. The Testbed is designed for physical layer research and experimentation following several emerging trends, such as cell-free networking, integrated communication, sensing, open disaggregated Radio Access Networks, AI-Native design, and multiband operation. The software is almost entirely based on free and open-source software, making contributing and reusing any component easy. The open Testbed is designed to provide real-time and labeled data on all parts of the physical layer, from raw IQ data to synchronization statistics, channel state information, or symbol/bit/packet error rates. Real-time labeled datasets can be collected by synchronizing the physical layer data logging with a positioning and motion capture system. One of the main goals of the design is to make it open and accessible to external users remotely. Most tests and data captures can easily be automated, and experiment code can be remotely deployed using standard containers (e.g., Docker or Podman). Finally, the paper describes how the Testbed can be used for our research on joint communication and sensing, over-the-air synchronization, distributed processing, and AI in the loop.

Building a real-time physical layer labeled data logging facility for 6G research

TL;DR

The architecture and vision of designing and implementing a new test infrastructure for 6G physical layer research at KU Leuven and how the Testbed can be used for research on joint communication and sensing, over-the-air synchronization, distributed processing, and AI in the loop are described.

Abstract

This work describes the architecture and vision of designing and implementing a new test infrastructure for 6G physical layer research at KU Leuven. The Testbed is designed for physical layer research and experimentation following several emerging trends, such as cell-free networking, integrated communication, sensing, open disaggregated Radio Access Networks, AI-Native design, and multiband operation. The software is almost entirely based on free and open-source software, making contributing and reusing any component easy. The open Testbed is designed to provide real-time and labeled data on all parts of the physical layer, from raw IQ data to synchronization statistics, channel state information, or symbol/bit/packet error rates. Real-time labeled datasets can be collected by synchronizing the physical layer data logging with a positioning and motion capture system. One of the main goals of the design is to make it open and accessible to external users remotely. Most tests and data captures can easily be automated, and experiment code can be remotely deployed using standard containers (e.g., Docker or Podman). Finally, the paper describes how the Testbed can be used for our research on joint communication and sensing, over-the-air synchronization, distributed processing, and AI in the loop.
Paper Structure (15 sections, 3 figures)

This paper contains 15 sections, 3 figures.

Figures (3)

  • Figure 1: The Testbed revolves around a central cluster of machines and USRP software-defined radios. The cluster comprises multiple machines: a gateway for remote access and container deployment, a controller to drive the USRPs and perform real-time signal processing, and a compute server for long-running machine learning tasks.
  • Figure 2: Different types of processing pipelines for O-RAN splits 7.2a and b. Depending on the vendors, stages can be distributed differently between O-RU and O-DU. vidaMaga
  • Figure 3: Basic software stack in use by the KU Leuven Cluster.