摘要
提出了一种用于在无线网络中传输视频的结合信源特性及网络拥塞控制的鲁棒性算法.通过场景建模以及特性分析,将分级编码产生的所有码流层划分成不同的类型,并根据它们对网络拥塞控制的贡献以及对重建图像质量的贡献不同,将其分成两个不同的队列.系统根据不同的网络丢包状态(即丢包是由网络拥塞引起还是由无线信道的不可靠传输引起)动态地调整信源速率、不等错误保护强度以及拥塞控制策略.仿真结果表明,该方法与MPEG-4信源编码加固定速率Turbo码方法以及动态调整信源、信道编码速率加选择性丢I,B,P包的网络拥塞控制方法相比,能够提供更好的性能.
An algorithm combining source statistics and network congestion control for robust video transmission over wireless network is proposed in this paper. Based on scene modeling and characteristic analyzing, all layers generated by scalable coding are classified into several types and two queuing are made respectively according to their contribution to network congestion and to the quality of reconstructed video. Then the source rate, unequal error protection level and congestion control strategy are dynamically adjusted according to different network status in which the packet loss is caused by network congestion or by unreliable transmission in wireless channel. The simulation results show that the proposed method can achieve better results than MPEG-4 video source coding with fixed rate Turbo coding and than the one which dynamically adjusting source coding and channel coding rates and using network congestion control method of selectively dropping L B or P packets.
出处
《软件学报》
EI
CSCD
北大核心
2007年第11期2882-2892,共11页
Journal of Software
基金
Nos.60702058,60532060(国家自然科学基金)
No.51434050105DZ0101(国防重点实验室基金)~~
