Development of A Novel Scheduling Algorithm for 5G RAN Using Deep Learning and Convolutional Neural Networks

Abstract

The fastest-growing segment of the communication sector and one of the most active research areas is wireless communication. Wireless data traffic has significantly increased as a result of the widespread use of mobile devices and the data-hungry apps that run on them. To handle these constantly rising data demands and transmission rates, the current cellular network architecture needs to be reevaluated. Device-to-device communications, which seem to be a possible solution to this issue, enable two mobile devices to connect with one another while they are in close proximity. Among the services it facilitates are messaging, video conferencing, data sharing, gaming, and content sharing. Increased spectrum efficiency, traffic unloading, system capacity, coverage area expansion, and throughput are just a few advantages of D2D communication. Widespread D2D communication deployment in diverse networks does, however, provide certain inherent challenges that must be addressed, including significant interference, mode selection, resource allocation, power control, security, and quality of service. Femtocell and D2D communication technology can greatly increase user throughput. The complex three-tier interference between macrocell-femtocell-D2D systems is a problem in heterogeneous networks, though. Interference is guaranteed in the system since D2D and cellular users share the same spectrum of resources during underlay D2D transmission. Therefore, careful coordination is required for spectrum resource allocation in D2D communication. The MATLAB simulation is used to validate the suggested model. The findings show that the suggested model lowers transmit power, produces Pareto-efficient outcomes, and boosts throughput in terms of UE rate and user satisfaction. Additionally, we showed that the suggested method converges to the Stackelberg game’s perfect subgame equilibrium