监测型传感网的“预测-修正”数据管理方法

"Predict-Revise" Data Management Method for Wireless Sensor Networks of Monitoring

将无线传感器网数据管理技术分为4类:陈述性语句、信号处理、模型预测以及中间件,并进行对比研究总结出如何既将传统的数据管理技术移植到传感网,又考虑到传感网自身的能源和网络带宽约束.针对监测型传感网,提出了一种新的"预测-修正"传感网数据管理方法,对传感节点的每个测量建立一个对应的虚拟表,所有的虚拟表又组成一个虚拟数据库.通过读取、预测、修正3个步骤,获得较为精确的测量值.最后用实验室温度监测实验验证了本方法.

The data management techniques of wireless sensor networks （WSN） are summarized into four categories,i, e. , declarative language,signal processing,model prediction,and middleware. This paper describes the key of moving the traditional data management techniques to WSN, and considers the constraints of power consumption and band width limit at the mean time. For the WSN of monitoring,a new ＂Predict-Revise＂ data management method is proposed. It creates one virtual table for each measurement variable of sensor nodes,and combines all the tables into a virtual database. After approaching reading, prediction,revise, the system will get the result with higher accuracy. An experiment for monitoring the temperature of lab is presented to verify this new method.