CAVIAR: Co-simulation of 6G Communications, 3D Scenarios and AI for Digital Twins

TL;DR

The paper tackles the need for accurate digital twins in 6G by proposing CAVIAR, a modular co-simulation framework that jointly couples wireless channel modeling, photorealistic 3D CGI, and AI/ML within a single loop. It introduces an orchestrator-driven, software-agnostic architecture and a message-passing mechanism to plug in diverse simulators, enabling all-in-loop interactions. The SAR blueprint demonstrates feasibility on a single workstation using AirSim/Unreal, Sionna, YOLOv8, and scikit-learn, coordinated via NATS, with benchmarking data on computational resource usage. The results highlight feasibility and scalability on limited hardware while identifying bottlenecks in the 3D ray-tracing path, guiding future work toward PTwin-in-the-loop and distributed co-simulation for broader 6G digital twin applications.

Abstract

Digital twins are an important technology for advancing mobile communications, specially in use cases that require simultaneously simulating the wireless channel, 3D scenes and machine learning. Aiming at providing a solution to this demand, this work describes a modular co-simulation methodology called CAVIAR. Here, CAVIAR is upgraded to support a message passing library and enable the virtual counterpart of a digital twin system using different 6G-related simulators. The main contributions of this work are the detailed description of different CAVIAR architectures, the implementation of this methodology to assess a 6G use case of UAV-based search and rescue mission (SAR), and the generation of benchmarking data about the computational resource usage. For executing the SAR co-simulation we adopt five open-source solutions: the physical and link level network simulator Sionna, the simulator for autonomous vehicles AirSim, scikit-learn for training a decision tree for MIMO beam selection, Yolov8 for the detection of rescue targets and NATS for message passing. Results for the implemented SAR use case suggest that the methodology can run in a single machine, with the main demanded resources being the CPU processing and the GPU memory.

Figures (12)

• Figure 1: Scenes for three distinct time instants of a CAVIAR simulation with an UAV being served via beamforming. The red pointy beam is shown for illustrative purposes, to indicate the UAV wireless channel was calculated via ray tracing.
• Figure 2: Representation of the three main CAVIAR modules and their interconnection via the orchestrator.
• Figure 3: CAVIAR simulation categories. (a) all-in-loop: three modules receive and send data in runtime; (b) AI/Comm-in-loop: AI and Communications inside the loop and (c) Mob+3D/Comm-in-loop: Mobility+3D and Communications within the main loop.
• Figure 4: Example of a publisher/subscriber dynamic within CAVIAR. In this case, the Mobility+3D module publishes UE Cartesian coordinates to subscribers through a given topic.
• Figure 5: The CAVIAR blueprint for the use case presented in this work.