Vision-Assisted Digital Twin Creation for mmWave Beam Management
This work addresses the problem of making digital twin technology more feasible for real-world mmWave communication networks, though it appears incremental as it builds on existing methods with a practical twist.
The paper tackles the challenge of creating accurate 3D digital twins for mmWave systems by proposing a pipeline and channel simulator that uses only a single camera and position information, demonstrating performance benefits in beam acquisition tasks on a real-world dataset.
In the context of communication networks, digital twin technology provides a means to replicate the radio frequency (RF) propagation environment as well as the system behaviour, allowing for a way to optimize the performance of a deployed system based on simulations. One of the key challenges in the application of Digital Twin technology to mmWave systems is the prevalent channel simulators' stringent requirements on the accuracy of the 3D Digital Twin, reducing the feasibility of the technology in real applications. We propose a practical Digital Twin creation pipeline and a channel simulator, that relies only on a single mounted camera and position information. We demonstrate the performance benefits compared to methods that do not explicitly model the 3D environment, on downstream sub-tasks in beam acquisition, using the real-world dataset of the DeepSense6G challenge