电子信息对抗技术
電子信息對抗技術
전자신식대항기술
ELECTRONIC INFORMATION WARFARE TECHNOLOGY
2015年
1期
39-42,92
,共5页
SVD%KNN%相似度%情报处理%变长信号序列%关联识别
SVD%KNN%相似度%情報處理%變長信號序列%關聯識彆
SVD%KNN%상사도%정보처리%변장신호서렬%관련식별
SVD( singular value decomposition)%KNN( K-nearest-neighbor)%similarity%intel-ligent process%signal sequences%identification
区别于传统的典型参数比对方法,将雷达信号理解为一种多特征变长数据来研究其关联识别。首先,应用奇异值分解方法得到雷达信号序列的特征向量;其次,基于特征向量对训练模板和测试数据进行相似度计算;最后,对每类训练数据采用滑窗提取模板以保证完备性和容错性,并用K近邻方法获取相似度最大的模板类别,从而实现多特征变长雷达信号序列的关联识别。仿真实验表明,该方法能以较高准确率实现雷达信号的关联识别,具有较强的普适性。
區彆于傳統的典型參數比對方法,將雷達信號理解為一種多特徵變長數據來研究其關聯識彆。首先,應用奇異值分解方法得到雷達信號序列的特徵嚮量;其次,基于特徵嚮量對訓練模闆和測試數據進行相似度計算;最後,對每類訓練數據採用滑窗提取模闆以保證完備性和容錯性,併用K近鄰方法穫取相似度最大的模闆類彆,從而實現多特徵變長雷達信號序列的關聯識彆。倣真實驗錶明,該方法能以較高準確率實現雷達信號的關聯識彆,具有較彊的普適性。
구별우전통적전형삼수비대방법,장뢰체신호리해위일충다특정변장수거래연구기관련식별。수선,응용기이치분해방법득도뢰체신호서렬적특정향량;기차,기우특정향량대훈련모판화측시수거진행상사도계산;최후,대매류훈련수거채용활창제취모판이보증완비성화용착성,병용K근린방법획취상사도최대적모판유별,종이실현다특정변장뢰체신호서렬적관련식별。방진실험표명,해방법능이교고준학솔실현뢰체신호적관련식별,구유교강적보괄성。
Different from all traditional parameter comparing methods, the radar signal is consid-ered as a type of variable length data of multi-attributes. Firstly, singular value decomposition ( SVD) properties of radar signal matrices are exploited to obtain the representative vector. Sec-ondly, similarities of pattern matrices and test matrices are calculated. Thirdly, patterns are ex-tracted by slipper window, and K-nearest-neighbor ( KNN) is applied to find the most similar pattern to the vector of test signal sequence. Simulations show that our method could achieve ap-proving result for radar signal identification, and could be widely applied.