基于神经网络的无缓冲数字系统延时控制模型  

Delay Control Model of Unbuffered Digital System Based on Neural Network

作　　者：应卫强[1,3] 罗仕鉴 张玲燕[3] YING Wei-qiang;LUO Shi-jian;ZHANG Ling-yan(City College,Zhejiang University,Hangzhou Zhejiang 310015,China;College of computer science and technology,Zhejiang University,Hangzhou Zhejiang 310058,China;School of Software Technology,Zhejiang University,Hangzhou Zhejiang 310058,China)

机构地区：[1]浙江大学城市学院,浙江杭州310015 [2]浙江大学计算机科学与技术学院,浙江杭州310058 [3]浙江大学软件学院,浙江杭州310058

出　　处：《计算机仿真》2021年第11期249-253,共5页Computer Simulation

基　　金：浙江省科技计划项目(2018C01086)。

摘　　要：传统的数字系统延时控制过程不能对系统的时间延迟进行准确预测,导致其存在拟合率低、网络传输延时过长和负载率低的问题。为解决上述问题,基于神经网络设计了新的无缓冲数字系统延时控制模型。使用时间戳技术记录采集到的数据包时间,并根据延时预测器预测在线网络延时。采用神经网络SMITH预估结果对无缓冲数字系统延时进行预估补偿,在SMITH估计模型的基础上构建无缓冲数字系统延时控制模型。通过PIP控制算法获得无缓冲数字系统延时控制模型的最优解。实验结果表明,上述模型的拟合率和负载率均较高,且能够有效减少网络传输延时,充分说明了上述模型的有效性。The traditional digital system delay control process can not accurately predict the system delay, leading to many disadvantages, such as low fitting rate, load rate and long network transmission delay. A new time-delay control model of unbuffered digital system based on neural network was designed in this paper. Timestamp technology was adopted to record the collected packet time, and the online network delay was predicted by the delay predictor. The neural network SMITH was applied to predict the results and compensate the unbuffered digital system. Based on SMITH′s estimation model, the delay control model of unbuffered digital system was founded. A PIP control algorithm was introduced to obtain the optimal solution of the delay control model of unbuffered digital system. The experimental results show that the model has high fitting rate and load rate, reduced network transmission delay and excellent effectiveness.

关 键 词：神经网络 延时控制 估计模型 控制算法 网络延时预测 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]