摘要： 针对VBR多媒体流量的数字特征，在混合网络流量预测模型的基础上，结合小波分析和卡尔曼滤波，提出一种多媒体流自适应带宽分配算法，以合理分配网络带宽，利用双重漏桶控制网络流量。在NS网络仿真平台上进行的仿真实验表明，该算法能有效避免网络拥塞，降低网络传输时延和数据丢包率，提高带宽利用率。

关键词: 预测, 拥塞控制, 服务质量

Abstract: According to the characters of MPEG Variable Bit Rate(VBR) traffic, this paper proposes a new network traffic prediction model combinated wavlet analysis and Kalman filter. A VBR traffic congestion management scheme based on the model is presented that is composed of Dual Leaky Bucket(DLB). Simulation on NS platform shows that the scheme can effectively avoid network congestion effectively, reduces the time-delay of network transmission and the lose packet rate, and improves the bandwidth usage.

Key words: prediction, congestion control, QoS

中图分类号: 

• TP393