无线多跳网络中SVC视频流传输的分布式优化方法  

Distributed Optimization Method for SVC Streaming in Wireless Multi-Hop Networks

作　　者：由磊[1] 雷建军[1] 王丕超[1] 

机构地区：[1]天津大学电子信息工程学院,天津300072

出　　处：《天津大学学报》2013年第1期73-78,共6页Journal of Tianjin University(Science and Technology)

基　　金：国家自然科学基金资助项目(61202380;61002029);天津市自然科学基金资助项目(12JCQNJC00300;12JCYBJC10400);教育部高等学校博士学科点专项科研基金(新教师类)资助项目(20110032120029);天津大学自主创新项目

摘　　要：针对无线多跳网络承载多媒体服务所面临的可扩展性、能量有效性和服务质量保证的挑战,研究了多个可伸缩编码(scalable video coding,SVC)视频流在无线多跳网络中传输的分布式跨层优化问题.在分析网络资源分配约束和SVC视频内容自适应模型的基础上,把SVC视频的传输问题表达为一个视频质量和网络功率消耗的跨层权衡优化问题.基于对偶问题的可分解性和次梯度求解方法,提出了一个可分布式实现的SVC视频优化传输算法.该分布式算法避免了控制开销的全网传播,通过节点的局部信息交换和独立解决相应子问题,达到底层资源分配和上层SVC视频传输内容选择的最优匹配.仿真结果表明该算法可以快速收敛到最优解,并能实现SVC视频的平均传输质量和网络总功率消耗的最优权衡.To meet the challenges of the scalability, energy efficiency and service quality guarantee for wireless multi-hop networks （WMN）in supporting multimedia, the distributed scalable video coding （SVC） streaming over such networks with cross-layer design is presented. A cross-layer trade-off optimization between the video quality and the power consumption of the WMN is formulated with constraints of the network resource （bandwidth and power） allocation and the adaptation model of SVC videos. Based on the decomposition and subgradient method of the dual problem, a cross-layer optimization algorithm for SVC streaming that can be implemented in a distributed man- ner is proposed. The distributed algorithm is capable of avoiding the diffusion of control overheads over the whole network, and achieve optimal matching of the resottrce allocation in the lower network layer and the content extrac- tion of the SVC videos through local information exchanging and solving independent subproblems in the relevant nodes. Numerical simulation results show that the algorithm is capable of converging quickly and obtaining optimal trade-off between the average received quality of the SVC steaming and the total power consumption of the WMN.

关 键 词：无线多跳网络 可伸缩编码 分布式优化 次梯度算法 

分 类 号：TN919.8[电子电信—通信与信息系统]