Digital Twins for Maintaining QoS in Programmable Vehicular Networks

Digital twins are virtual replicas of the real world that capture precise interactions between actual devices and the environment in which they operate. Enabled by advancements in computing and ubiquitous software, digital twins may receive data from the real world as well as generate virtual datasets for making decisions that ultimately influence the complex real-world actions. As a use case, we propose applying digital twins in connected and autonomous vehicles for exceeding the wireless quality of service (QoS) in non-line-of-sight (NLOS) scenarios. Key to this vision is incorporating programmable base stations (BS) and reconfigurable intelligent surfaces (RIS) that mitigate the frequent blockages in dynamic vehicular networks. In our method, a LiDAR sensor mounted autonomous car detects obstacles between itself and the base station (BS) it is connected to during its voyage in order to predict NLOS situations, which are then reported to the BS, where the digital twin is deployed. Upon an obstacle notice, the digital twin is triggered and runs ray-tracing by placing an obstacle in the digital twin's virtual map in order to determine BS settings, such as transmit power, antenna directivity, wireless standard, and RIS configurations that will maintain the QoS along the autonomous vehicle's trajectory. We validate the feasibility of our proposed method using real life LiDAR data and Wireless InSite ray-tracing software, in which we determine the base-station and RIS configurations to maintain QoS by testing wireless standards, that is, 802.11ac and LTE, antenna patterns and directivity, and transmit power levels. Results reveal that can detect obstacles from 11 m away, giving enough time for BS to reconfigure itself and RIS within city speed limits, and reduce the required transmit power by 2.32x, on average in dBm, compared to the LOS scenarios through RIS.