无人机故障特征提取问题研究  

Research on UAVS’Fault Feature Extraction

作　　者：谭勇[1] 赵铭[1] 

机构地区：[1]电子工程学院,合肥230037

出　　处：《航空电子技术》2010年第1期45-48,53,共5页Avionics Technology

摘　　要：针对无人机故障诊断前故障特征提取的问题进行了分析。引入了主成分分析法,用于对初始的特征参数样本进行降维处理。根据无人机特征参数样本复杂多维的特点,结合故障诊断时对特征向量相关性的要求,提出按功能模块对特征样本进行分类的改进方法,达到初步降维和保存需要的相关性向量的要求。In this article, the issue of extraction of fault features in UAV is analyzed. By means of primary component analysis, the. dimension of initial characteristic, parameter samples is significantly reduced. According to the complication of UAV＇s characteristic parameter samples and multi-dimension, as well as requirements of eigenvector＇s dependency in fault diagnosis, an improved feature selection method used to classify the characteristic parameter samples in accordance with the functionality modules is proposed, This method can reduce the dimensions while a certain set of eigenvector can remain dependent.

关 键 词：无人机 特征提取 主成分分析 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]