摘要： 根据网络流量的变化规律，提出自回归模型下的抖动缓冲控制算法。通过改进的随机中点置位算法，建立具有突发性和自相似 性的网络业务流量预测模型。依据网络流量的预测值设置缓冲区大小，并在使用中不断改进以提高缓冲区设置精度。采用Matlab软件对抖动缓冲控制算法进行仿真，基于E-modle对语音质量进行评估，结果表明该算法在没有其他服务质量保证的情况下，MOS值均在中级标准以上，较好地改善了网络电话的语音质量。
关键词:
时延抖动,
缓冲算法,
统计特性,
自回归模型,
自相似性
Abstract: A jitter buffering control algorithm based on Auto Regression(AR) model is proposed according to the changing regularity of network flux. A network business flux prediction model with burstiness and self-similarity is built by an improved RMD arithmetic. It sets buffer area size based on the prediction value of network flux, and improves the buffer setting precision. Result of assessment to Quality of Service(QoS) of Voice over Internet Protocol(VoIP) and simulation of jitter buffering control algorithm by using Matlab software shows that the MOS value outscores secondary level, and the mean value reaches a better level, the QoS of VoIP is improved.
Key words:
delay jitter,
buffering algorithm,
statistical property,
Auto Regression(AR) model,
self-similarity
中图分类号:
赵国栋, 李栓红, 黄国策, 郑博. 基于网络流量预测的抖动缓冲控制算法[J]. 计算机工程, 2012, 38(7): 80-82,86.
DIAO Guo-Dong, LI Shuan-Gong, HUANG Guo-Ce, ZHENG Bo. Jitter Buffering Control Algorithm Based on Network Flux Prediction[J]. Computer Engineering, 2012, 38(7): 80-82,86.