Proof-of-Concept of Network Digital Twin-Based Channel Prediction for mmWave V2X Communications

Abstract

Millimeter-wave vehicle-to-everything (V2X) communications are a promising technology for supporting high-datarate cooperative sensing applications, such as LiDAR point-cloud sharing among vehicles and roadside infrastructure. However, mmWave links are highly sensitive to mobility-induced blockage, line-of-sight (LoS) transitions, and antenna directionality, which can cause rapid fluctuations in received signal strength and degrade the reliability of data delivery. To address this issue, this work presents a real-time Network digital twin (Network DT)-based received signal strength indicator (RSSI) prediction framework for mmWave V2X communications. The proposed framework couples a Mobility digital twin (Mobility DT), which provides live and predicted vehicle states, with a ray-tracingbased Network DT that estimates future RSSI values for direct V2V and infrastructure-assisted V2I2V links. We implement the framework in an outdoor proof-of-concept testbed using 60 GHz WiGig links and evaluate its prediction accuracy by comparing predicted RSSI with field measurements. Experimental results show that the proposed Network DT captures the temporal variation of measured RSSI with a mean absolute error of 2.19 dB and a root mean square error of 2.93 dB. Furthermore, a pointcloud transmission experiment demonstrates that the predicted RSSI provides useful link-quality information for assessing the feasibility of cooperative sensing data delivery over direct and roadside unit (RSU)-assisted paths. These results indicate that real-time Network DT-based RSSI prediction can serve as a practical foundation for proactive and reliable mmWave V2X communication control.