数据融合模型在专家诊断系统中的应用  

Application of data fusion model in expert diagnostic system

作　　者：王鹏[1] 冯泽安 Wang Peng;Feng Ze＇an(School of Electronic Information Engineering,Xi＇an Tenchnogical University,Xi＇an 710021,China)

机构地区：[1]西安工业大学电子信息工程学院

出　　处：《国外电子测量技术》2018年第9期120-125,共6页Foreign Electronic Measurement Technology

基　　金：陕西省科技计划重大重点项目(2017ZDXM-GY-114)资助

摘　　要：为了消除特种车辆专家诊断系统中无线传感器数据冗余和解决其寿命较短的问题,结合传感器的分布特点,将多传感器数据融合模型添加到系统中,先对无线传感器网络使用LEACH路由协议进行分簇,然后在此基础上应用深度神经网络（DNN）,对各簇中数据进行特征提取并融合处理;经仿真实验表明：该模型不但非常有效去除无效和冗余的数据,降低了传感器的的功耗,使其存活时间延长了一倍,同时数据复现动态误差降低到5.04%,达到了用少量数据反映出原始数据特征。In order to eliminate the problem of redundancy of wireless sensor data and solve the problem of short lifetime in special vehicle expert diagnosis system,the multisensor data fusion model is added to the system based on the distribution characteristics of the sensor.The LEACH routing protocol is used to divide the wireless sensor network.Clusters,and then use this depth neural network（DNN）,feature extraction and fusion processing of the data in each cluster;simulation experiments show that：this model is not only very effective to remove invalid and redundant data,reducing the sensor′s power consumption has doubled its survival time,and the dynamic error of data replication has decreased to5.04%,which has achieved a small amount of data to reflect the characteristics of the original data.

关 键 词：专家诊断系统 数据融合 LEACH路由协议 深度神经网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]