一种运用SVM分类的多源态势数据关联方法  

A Multisource Situation Data Association Method Based on SVM Classification

作　　者：李川[1] 黄威 孙伟[1] 郝欢 LI Chuan;HUANG Wei;SUN Wei;HAO Huan(Unit 31101 of PLA,Nanjing 210014,China)

机构地区：[1]中国人民解放军31101部队,南京210014

出　　处：《电讯技术》2022年第12期1809-1812,共4页Telecommunication Engineering

基　　金：国家社会科学基金军事学项目(2020-SKJJ-A-C-001)。

摘　　要：阐述如何利用多手段获取的战场多源目标态势数据生成战场态势一张图,进一步提高海上方向的监控能力。针对多源目标态势数据存在的时空误差、属性模糊等问题,提出一种运用支持向量机(Support Vector Machine,SVM)分类的多源态势数据关联方法。运用机器训练学习方式,获取不同误差下的多源态势关联,识别非线性特征参数,构建多源态势数据关联算法模型,有效形成目标批号和时空位置的唯一性。通过模拟真实数据进行验证,结果表明该算法实现效果较好,在多源态势融合方面具有较好的实用价值。This paper discusses how to use the battlefield multi-source target situation data received by multiple means to generate a map of the battlefield situation and further improve the ability of maritime surveillance.For the multi-source target situation data,which exists some questions with space-time errors and fuzzy attributes,an association method for multisource situation data based on support vector machine(SVM)classification is proposed.By using machine-learning algorithm,the multisource situation association under different errors is obtained,the nonlinear characteristic parameters are identified,the algorithm model of multisource situation data association is constructed,thus forming the uniqueness of target batch number and space-time position effectively.In addition,by simulating the real data,the results show that the algorithm has good implementation effect and practical value in aspect of multisource situation fusion.

关 键 词：多源态势融合 多源态势关联 机器训练学习 支持向量机(SVM) 特征样本 

分 类 号：TN97[电子电信—信号与信息处理] TP391[电子电信—信息与通信工程]