基于支持向量机的机内测试降虚警技术  

The BIT False Alarm Reducing Technology Based on Support Vector Machine

作　　者：徐哲[1] 刘冠军[1] 邱静[1] 吕克洪[1] 

机构地区：[1]国防科技大学机电工程与自动化学院,湖南长沙410073

出　　处：《测试技术学报》2006年第5期424-428,共5页Journal of Test and Measurement Technology

摘　　要：针对环境因素引起的B IT虚警,提出基于支持向量机的机内测试降虚警技术,建立了虚警与环境因素的关联关系.通过构建多类SVM分类器,将间歇故障与正常状态和硬故障分开,以降低B IT虚警.某型号航空地平仪的试验结果分析表明:在小训练样本的情况下,支持向量机比神经网络具有更好的间歇故障识别能力,能够有效地降低由环境因素引起的虚警.Focused on the built-in test （BIT） false alarm caused by environmental factor, a BIT false alarm reducing technology based on support vector machine （SVM） is proposed. The dependence relation between the false alarm and the environmental factor is established by support vector machine. A multiple SVM classification is designed to distinguish the intermittent failure from the normal and the hard failure.. According to the experimental result of an altazimuth, it has been proved that Support Vector Machine is superior to Neural Network in the recognition of the intermittent failure under the small training samples, and the BIT false alarm caused by＇ environmental factor can be effectively reduced by this means.

关 键 词：机内测试 虚警 支持向量机 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]