STAR-RIS Based Resource Scheduling and Mode Selection for Drone Assisted 5G Communications

For a range of future vehicle usage in 5G, communication between vehicle road cooperative systems (VRCS) is cru-cial. With the advent of 5G deployments, vehicle-to-everything (V2X) communications is enabled to boost network density, optimize transmission mode selection, expand resource capacity, and provide near-ubiquitous connectivity between vehicles with extremely dependable and low latency. On the other hand, mode selection and resource scheduling in V2X communications come with intrinsic difficulties. In this paper, we present a joint optimization approach to address the problems associated with scheduling resources and choosing a transmission mode for the drone-mounted Simultaneous Transmitting and Reflecting-Intelligent Reflecting Surface (STAR- RIS) infrastructure for virtual reality surveillance. The Markov decision process is used to formulate the problem, and the deep reinforcement learning (DRL)-based STAR-RIS drone mounted approach maximizes vehicle-to-infrastructure user capacity while meeting the latency and reliability requirements of vehicle-to-vehicle (V2V) pairs. Moreover, a DRL approach called PDQN (Parametrised Deep Q Network) is employed to generate resilient models while considering the training constraints of local D RL models. The results of the simulation validate the PDQN DRL algorithm's superiority for V2V pairs and show that the proposed DRL-based algorithm outperforms current baseline techniques.