Digital twin (DT) technologies have been increasingly important and useful for wireless communications. In particular, to support soaring wireless traffic with limited frequency spectrum assigned to legacy cellular systems, efficient operation and management of wireless resources as well as preemptively prediction of future spectrum usage are crucially important. For such purpose, DT networks for current fourth-generation (4G) and fifth-generation (5G) networks are constructed in this paper, by simultaneously utilizing measurement data from user equipments (UEs) and geographical information and characteristics of 4G and 5G base stations (BSs) within a specific observation area. Representative case studies are provided to demonstrate the usefulness of DT enabled cellular network management and prediction. As a real or near real time DT application, the impact and benefit of dual connectivity and dynamic spectrum sharing between 4G and 5G networks are analyzed. As a non-real time DT application, long-term improvement of 5G networks such as densification of BSs, implementation of advanced multiple input and multiple output technologies, and assignment of additional spectrum are analyzed and compared.
