Digital twin creates a digital representation of the physical world, which enables immersive technologies such as augmented reality (AR), virtual reality (VR), and holographic communication. In this paper, we take a significant step towards enabling digital twin at scale & low-cost by estimating depth map of the environment using existing communication infrastructure. Traditionally, camera or LiDAR (light detection and ranging) have been used for estimating the 3-D depth map. The proposed method RF2LiDAR is first of its kind algorithm that generates LiDAR-like high resolution representation of the environment from ambient communication signals. We first perform preprocessing on the multiple-input-multiple-output (MIMO) radio-frequency (RF) signal and then input the processed data to a deep learning model to target the LiDAR point cloud data. RF2LiDAR is able to generate LiDAR-like depth map of room of size 19m × 10m × 2m with 0.25m granularity from MIMO data. Further, we show that the predicted point clouds have an average Chamfer distance of 1.5m2 and they capture the change in perception across various testing locations without any prior information of location and orientation of the receiver.
RF2LiDAR: Enabling Digital Twin Using MIMO RF Signals
24.06.2024
2501653 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
On the modeling of twin rotor mimo system using chirp inputs as test signals
British Library Online Contents | 2017
|A Twin-Cluster MIMO Channel Model
British Library Conference Proceedings | 2006
|