Resource Optimization For Video Streaming Multicast Services In UAV-Assisted Intelligent Transportation Systems

Unmanned Aerial Vehicle(UAV)assisted Intelligent Transport Systems(ITS)serve as a new paradigm for smart cities,providing significant opportunities for traffic video applications.UAVs equipped with high-definition cameras can send a comprehensive view of traffic to Connected Autonomous Vehicles(CAVs),assisting them in traffic guidance and traffic activity analysis.However,due to the high resolution,frame rate,and dynamic range characteristics of real-time high-definition videos,a large amount of wireless resources are required to send real-time high-definition videos to vehicles.Therefore,optimizing resources to improve the Quality of Experience(QoE)for vehicle users is a significant challenge in UAV-assisted ITS.This paper studies the resource optimization of video multicast services in UAV-assisted ITS,aiming to maximize the long-term video quality received by all vehicles.The specific work are as follows:(1)This article proposes a research scheme for optimizing traffic video multicast service resources with the assistance of Unmanned Aerial Vehicles in Intelligent Transportation Systems.The scheme utilizes stream decomposition(scalable video coding)and group decomposition techniques to maximize the long-term video quality received by all vehicles through a joint optimization of vehicle grouping and Physical Resource Blocks(PRBs)allocation strategies.Considering the mutual interference between UAVs,the above joint optimization problem is first modeled as a Markov Game(MG)process.Then,the state-of-the-art Multi-Actor Attention Critic algorithm is employed to solve the MG,which uses attention mechanisms to make the learning process more effective and scalable.Finally,simulation results show that the Multi-Actor Attention Critic algorithm performs well in terms of profit performance,convergence stability,and learning efficiency.(2)This article proposes a research scheme for optimizing traffic video multicast service resources with the assistance of Unmanned Aerial Vehicle in Non-Orthogonal Multiple Access(NOMA)-Scalable Video Coding(SVC)in Intelligent Transportation Systems(ITS).The SVC multicast in this scheme uses NOMA technology to maximize the long-term video quality received by all vehicles through a joint optimization of vehicle grouping and power allocation strategies.The above joint optimization problem is first modeled as a Markov Decision Process(MDP).Then,the Soft Actor-Critic algorithm is employed to solve it through deep reinforcement learning.Finally,simulation results show that the Soft Actor-Critic algorithm has better convergence stability and profit performance than the traditional Actor-Critic algorithm.