Cross-Layer Optimization Techniques For Wireless Multimedia Communication

With the rapid development of mobile communication and ubiquitous deployment of wireless networks,the mobile video has experienced an unprecedented growth and the wireless multimedia transmission technology has become a research hotspot.However,the deficiencies of traditional video streaming framework appear gradually.The main reason is the corresponding video compression part and channel coding part are separated,which is proven to be optimal only for the point to point communication.And this separated scheme can not efficiently allocate the resources to multiple users and guarantee the quality of their received video.To resolve the problem,the uncoded wireless video streaming system is proposed.It uses cross-layer approach to jointly optimize the video compression and channel coding.Since only linear operations are applied in the entire process,the quality of received video is proportional to the channel condition.Here,the dissertation also consider the non-orthogonal multiple access technique and its application in the video streaming framework.It uses signal multiplexing and joint resource allocation for multiple users to enhance the system performance.Though the current works on cross-layer optimization techniques for wireless multimedia communication has made some progresses,the drawbacks still exist.First,due to the wireless channels fluctuate over time in reality,the received quality of users may not be good enough.Second,the multiple users will be in different channel conditions and compete for transmission resources.This makes the corresponding resource allocation problem for multiple users highly complex and hard to be solved,especially when considering to apply the non-orthogonal multiple access technique into video streaming system.Therefore,designing the new uncoded wireless video streaming system and multiuser resource allocation scheme that exploits the cross-layer optimization for video compression and channel coding can enhance the efficiency of resources utilization,which is the key to improve the system performance of wireless multimedia communication.In this dissertation,the problem analysis,slolution reserach and numerical simulation are carried out on the design of wireless multimedia communication system.The objective is to improve the current solution and realize high-quality wireless video transmission.The specific research works are listed as follows.(1)An uncoded wireless video transmission system that utilizes time and frequency diversities in multiple time slots is proposed.It can realize the optimization on resource allocation and improve the quality of video received by user.(2)An uncoded multiuser video streaming system that exploits diversities of video contents and channel conditions of multiple users is proposed.It reasonably allocates resource to multiple users and can improve the mean-square error performance of users.(3)An uncoded NOMA-based multiuser video streaming system is proposed to resolve the multiuser signal multiplexing and joint resource allocation problem.It can also achieve the high-quality multiuser video transmission.(4)A episodic deep reinforcement learning framework is proposed to solve the joint channel assignment and power allocation problem with high complexity in NOMA system.It can also achieve near optimal system performance.