基于多项式预测滤波理论的实时信号传输  

Real-time signals transmission based on polynomial predictive filtering theory

作　　者：董劲男[1] 秦贵和[1] 张晋东[1] 崔玥[1] 

机构地区：[1]吉林大学计算机科学与技术学院,长春130012

出　　处：《吉林大学学报（工学版）》2008年第4期890-896,共7页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家发展改革委员会下一代互联网示范工程2005年研究开发产业化及应用试验项目(3H006l421421)

摘　　要：利用信号处理理论研究实时信号网络传输问题,提出通过预测的方法来改善由于网络不确定性造成的不良后果。将多项式预测滤波理论与信道编码技术相结合构建一个看似从本地传感器输出信号的虚拟传感器,从而将网络传输的不确定性转化为系统模型的不确定性,并在此基础上将网络不确定性造成的误差和预测误差一并作为虚拟传感器的输出误差来处理,兼顾了信号传输的实时性和可靠性。经上述处理后,研究者可以像在非网络环境下一样分析和设计与实时信号网络传输有关的系统,本文方法对网络控制系统和传感器网络的发展具有重要理论意义和应用价值。The real-time signal transmission was studied by using signal processing theory. A prediction method to alleviate the bad consequences caused by the transmission uncertainty was presented. Between users and network a virtual sensor, whose output signals seem to be generated from a local sensor, was constructed by combining polynomial predictive filtering （PPT） theory and channel coding technique. Thus, the network transmission uncertainty was transformed into the system model uncertainty. Based on this, the error caused by the transmission uncertainties and the prediction error were processed together as the output error of virtual sensor, by which the real-time performance and the reliability of the signal transmission were balanced. In such a way, a system related with real-time signal transmission can be designed and analyzed as the non-networked system. The presented method is of great theoretical importance and application significance for the development of networked control system （NCS） and sensor network.

关 键 词：计算机应用 多项式预测滤波 采样 虚拟传感器 实时信号 网络传输 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]