| 摘要: |
| 随着敌我识别侦察装备的广泛部署，其侦察数据呈现爆炸式增长，多源且复杂的海量信号侦察数据对后端数据分析提出了较高要求。针对传统敌我识别信号关联分析方法速度慢、效率低的问题，通过研究其关联流程的可拆分特性，提出了一种基于Map/Reduce的敌我识别信号聚类和关联分析方法，将传统敌我识别信号关联分析流程进行方位和时间拆解，在大数据平台上进行分布式并行计算，大幅提升了大数据量情况下的关联分析效率。实验结果表明，在数据量达到108条以上时，所提方法比传统方法计算效率高10倍以上。 |
| 关键词: 敌我识别 大数据 信号聚类 信号关联 |
| DOI：10.3969/j.issn.1001-893x.2017.11.008 |
|
| 基金项目: |
|
| A Map/Reduce-based parallelization IFF signal correlation method |
| LIU Yong |
| () |
| Abstract: |
| With the wide deployment of identification between friend or foe (IFF) reconnaissance equipment, its reconnaissance data has exploded, mass multi-source and complex signal reconnaissance data puts forward higher requirements for the analysis of back end data. As the analysis of traditional stand-lone IFF signal correlation method is slow and inefficient, through researching on the split characteristic of the correlation process, this paper presents a new IFF signal clustering and correlation analysis method based on Map/Reduce, which dismantles the traditional method by azimuth and time and computes in distributed parallel way on large data platform, thus greatly enhancing the analysis efficiency. The experimental results show that the calculation efficiency is 10 times higher than that of the traditional method when the data volume reaches more than 108. |
| Key words: IFF big data signal clustering signal correlation |
《电讯技术》编辑部 