Abstract: In this paper, an adaptive link-level Automatic Repeat reQuest (ARQ) algorithm was proposed for wireless media streaming, in the hope that the media could be played on receivers with high quality. The algorithm, based on cross-layer optimization design, predicts the packet loss rate using the model of GM (1, 1), and adjusts the ARQ parameter Nmax according to the characteristics of the number of frames of a GOP. Besides, an adaptive FEC at the application layer was used to adjust the sending rate of MPEG video frame adaptively and to allocate the network bandwidth resource between the MPEG video source data and the redundant data dynamically. Both the mathematical analysis and NS simulation results demonstrate that the proposed mechanism can significantly improve quality of media streaming, in terms of playable frame rate, reliability and real-time performance on the receiving side.

Key words: wireless real-time media streaming, cross-layer design, GM (1, 1), adaptive link-level ARQ

摘要： 提出了一种用于无线实时流媒体传输的自适应链路层ARQ控制策略，用以提高接收方的播放质量。该策略采用跨层设计的方法，基于GM（1，1）模型预测当前的网络状态，考虑GOP可解码帧数的特性，自适应地调整ARQ参数Nmax；另一方面，在应用层采用自适应FEC策略，在视频源数据和冗余数据之间动态分配网络带宽。数学分析和仿真验证均表明，该策略能使接收方获得最大的可播放帧率，有效地提高了流媒体传输的可靠性和实时性。

关键词: 无线实时流媒体, 跨层设计, GM(1, 模型, 自适应链路层ARQ