【摘要】：雷達輻射源信號分選是現(xiàn)代電子對抗和未來信息戰(zhàn)中的首要技術(shù),也是電子情報偵察系統(tǒng)和電子支援系統(tǒng)中的瓶頸技術(shù)。只有雷達偵察設(shè)備正確分選的基礎(chǔ)上,才能有效識別輻射源信號,判定被截獲雷達的類型和威脅等級,進而選擇合適的對抗策略。隨著電子戰(zhàn)、信息戰(zhàn)的激烈對抗和現(xiàn)代雷達技術(shù)的快速發(fā)展,雷達信號分選所面臨的電磁環(huán)境日益復(fù)雜密集,采用低截獲概率、脈內(nèi)波形變換、多參數(shù)捷變等技術(shù),具有多種工作模式的先進體制雷達也在逐漸取代信號形式簡單、信號參數(shù)不變或只是緩慢變化的傳統(tǒng)雷達。針對復(fù)雜密集電磁環(huán)境下先進體制多模雷達輻射源信號處理中存在的關(guān)鍵理論問題,本論文進行了探索性研究,獲得如下研究成果:(1)分析多模雷達輻射源信號特征的基礎(chǔ)上,討論傳統(tǒng)未知輻射源信號分選模型處理多模雷達信號存在的問題,提出一種新的針對復(fù)雜電磁環(huán)境下未知多模雷達輻射源信號的分選模型,以更有效利用截獲雷達脈沖序列。(2)將數(shù)據(jù)挖掘理論引入到雷達信號分選技術(shù),通過數(shù)據(jù)場理論描述雷達輻射源信號樣本以抑制復(fù)雜電磁環(huán)境中高強度的噪聲干擾和離群數(shù)據(jù)干擾,根據(jù)數(shù)據(jù)場等勢線的分布情況建立信號樣本嵌套結(jié)構(gòu)供后續(xù)的層次聚類多參數(shù)分選算法使用,規(guī)避需要輻射源信號先驗知識才能完成的參數(shù)設(shè)定環(huán)節(jié)。(3)提出利用層次聚類算法進行多模雷達輻射源信號分選,層次聚類通過層次構(gòu)架模式,遞歸地對信號樣本進行合并或分裂,最終形成一種嵌套的類層次結(jié)構(gòu)或類譜系圖,進而將屬于同一輻射源不同工作模式的雷達信號歸納為同一譜系,可有效減少多模雷達信號分選過程中“增批”現(xiàn)象的出現(xiàn)。(4)利用云模型理論可以從模糊、隨機、不確定的小樣本數(shù)據(jù)中提取出定性概念的固有特征,提出一種基于云模型理論的分選結(jié)果有效性評估算法。算法中將每個層次聚類分選結(jié)果視為一個云模型,根據(jù)提出的評價判定準(zhǔn)則比較、處理不同云模型間的隸屬度。該算法可有效解決聚類算法類間分離度較差的問題,同時可反饋優(yōu)化多參數(shù)聚類分選過程。
[Abstract]:Radar emitter signal sorting is the most important technology in modern electronic countermeasure and future information warfare, as well as the bottleneck technology in electronic intelligence reconnaissance system and electronic support system. Only on the basis of the correct sorting of radar reconnaissance equipment can the emitter signal be effectively identified, the type and threat level of the intercepted radar can be determined, and the appropriate countermeasures can be selected. With the fierce confrontation of electronic warfare, information warfare and the rapid development of modern radar technology, the electromagnetic environment of radar signal sorting is becoming more and more complex and dense. The techniques of low probability of interception, in-pulse waveform transformation, multi-parameter agility and so on are used. The advanced system radar with various working modes is gradually replacing the traditional radar with simple signal form, constant signal parameters or only slowly changing signal parameters. In view of the key theoretical problems existing in the signal processing of advanced multi-mode radar emitter in complex and dense electromagnetic environment, this paper carries out an exploratory study. The main results are as follows: (1) on the basis of analyzing the characteristics of multi-mode radar emitter signal, the problems of traditional unknown-emitter signal sorting model for multimode radar signal processing are discussed. A new sorting model for unknown multi-mode radar emitter signals in complex electromagnetic environment is proposed in order to make more effective use of intercepted radar pulse sequences. (2) data mining theory is introduced into radar signal sorting technology. Radar emitter signal samples are described by data field theory to suppress high intensity noise interference and outlier data interference in complex electromagnetic environment. According to the distribution of the isopotential lines of the data field, the nested structure of the signal samples is established for the subsequent hierarchical clustering multi-parameter sorting algorithm. In order to avoid the parameter setting process which requires prior knowledge of emitter signal, the hierarchical clustering algorithm is proposed for multi-mode radar emitter signal sorting. The signal samples are merged or split recursively to form a nested class hierarchy or pedigree diagram, and then the radar signals belonging to different working modes of the same emitter are grouped into the same spectrum. It can effectively reduce the phenomenon of "increasing batch" in the process of multi-mode radar signal sorting. (4) using cloud model theory, the inherent characteristics of qualitative concepts can be extracted from fuzzy, random and uncertain small sample data. An algorithm for evaluating the validity of sorting results based on cloud model theory is proposed. In the algorithm, the clustering and sorting results of each level are regarded as a cloud model, and the membership degree among different cloud models is dealt with according to the comparison of the evaluation criteria proposed. The algorithm can effectively solve the problem of poor separation between clusters and can be used to optimize multi-parameter clustering process.
【學(xué)位授予單位】：哈爾濱工程大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TN957.51

【參考文獻】

相關(guān)期刊論文 前10條

1 邱超凡;李浩;;基于DAGSVM的雷達輻射源信號分選與識別[J];雷達科學(xué)與技術(shù);2011年03期

2 趙玉;陸志宏;;一種多模雷達信號分選方法的研究[J];現(xiàn)代電子技術(shù);2010年13期

3 朱崢嶸;;國外雷達技術(shù)新進展概述[J];信息化研究;2010年06期

4 劉慶偉;孫靜;;關(guān)系數(shù)據(jù)庫中多維數(shù)據(jù)分析及展現(xiàn)的研究[J];計算機工程與設(shè)計;2008年09期

5 國強;王常虹;郭立民;李崢;祝正威;華云;;分段聚類在雷達信號分選中的應(yīng)用[J];北京郵電大學(xué)學(xué)報;2008年02期

6 陳黎飛;姜青山;王聲瑞;;基于層次劃分的最佳聚類數(shù)確定方法[J];軟件學(xué)報;2008年01期

7 朱家昌;趙玉潔;賁德;;機載有源相控陣火控雷達發(fā)展現(xiàn)狀及趨勢[J];電訊技術(shù);2007年04期

8 宮鳳強;李夕兵;;距離判別分析法在巖體質(zhì)量等級分類中的應(yīng)用[J];巖石力學(xué)與工程學(xué)報;2007年01期

9 董光焰;鄭永超;張文平;王治中;吳登喜;趙明福;田勇;;相控陣激光雷達技術(shù)[J];紅外與激光工程;2006年S3期

10 邱保志;沈鈞毅;;基于擴展和網(wǎng)格的多密度聚類算法[J];控制與決策;2006年09期

相關(guān)博士學(xué)位論文 前1條

1 國強;復(fù)雜環(huán)境下未知雷達輻射源信號分選的理論研究[D];哈爾濱工程大學(xué);2007年

，

本文編號：2400722