基于群机并行的数据融合算法设计与实现

Design and Implementation of Data Fusion Algorithm Based on Cluster of Workstations(COW)or Network of Workstations(NOW)

目前综合电子战的情报侦察数据融合系统,一般只能适应中、小批次的平台,如果战场环境复杂,平台数目增加,其实时性便很难达到。网络并行是目前解决大型计算的一条方便、经济的道路。文章针对数据融合对辐射源、平台的识别推理需要较多时间这一问题,设计和实现了其基于群机并行的识别算法,并在群机NCICluster2000上成功运行,最后给出在其上的实验结果和分析。

Now Information Reconnaissance and Data Fusion Systems about comprehensive EW does not adapt to large scales with lots of batches of plates.If modern battlefield becomes more complicated and the number of plates increases,its real-time performance will decrease.COW(or NOW)is an economical and easy way to adapt to computing on a large scale.Since it needs much time for Data Fusion to recognize and deduce radiant points and plates,this paper designs and implements a method based on COW,which runs successfully on NCI Cluster 2000.Lastly,results and explains are given here in details.