Digital Twin of a Beam Selection Procedure in an Indoor Scenario

Millimeter waves (mmWave) communication, vital for 5G/6G, addresses sub-6 GHz congestion but needs directional beamforming due to high path loss. Machine learning (ML) for beam selection faces scarce indoor/near-field training data limitations. Digital Twin (DT) frameworks mitigate this by generating realistic datasets, often validated by simplified digital model (DM)s. This work presents a validated methodology to construct accurate DMs from low-cost radio measurements. An indoor mmWave multiple input single output (MISO) DM, using Wireless Insite (WI) ray tracing (RT), achieved 100% top-1 beam accuracy and strong received signal strength (RSS) correlation in Boresight line-of-sight (LoS) condition.