无线视频通信中基于队列预测的自适应缓存管理算法  

An Adaptive Buffer Management Algorithm Based on Queue Prediction for Wireless Video Communication

作　　者：钱超[1] 朱晓荣[1,2] 

机构地区：[1]南京邮电大学江苏省无线通信重点实验室,南京210003 [2]东南大学移动通信国家重点实验室,南京210096

出　　处：《电信科学》2014年第7期70-76,共7页Telecommunications Science

基　　金：国家自然科学基金资助项目(No.61372125);国家重点基础研究发展计划("973"计划)基金资助项目(No.2013CB329104);江苏省自然科学基金资助项目(No.BK2012831);东南大学移动通信国家重点实验室开放研究基金资助项目(No.2013D01)

摘　　要：无线网络中实时视频业务对网络环境的变化非常敏感,因此通常需要在接收端采用合理的缓存管理策略来缓解网络波动对用户观看视频造成的影响。提出一种基于队列预测的自适应缓存播放管理机制,该方法通过判断当前缓存队列的状态,并根据实时视频到达率和端到端时延对缓存长度以及播放速率进行综合调整。实验结果表明,所提算法可以随着网络环境的波动自适应地调整缓存大小和播放速率,有效降低视频业务的中断频率和丢帧率。The real-time video services in wireless networks are fragile to the dynamic network environment. Therefore, proper buffering management techniques are adopted to mitigate the impact of network fluctuation on user experiences. An adaptive buffer management algorithm based on queue prediction was proposed for wireless video streaming applications. By judging the current state of the buffer, the scheme adjusted the video playout rate and buffer size according to the real-time frame arrival rate and end-to-end delay. The experiments were conducted to validate the proposed algorithm. The experimental results show that this method is able to adaptively adjust buffer size and playout rate according to network fluctuations, which reduces the interrupt frequency and the number of lost frame effectively.

关 键 词：自适应缓存管理 队列预测 缓存状态 时延控制 

分 类 号：TN919.8[电子电信—通信与信息系统]