雷达有源欺骗干扰多尺度特征级识别技术研究

发布时间：2018-08-28 06:31

【摘要】：伴随着电子干扰技术的发展,更加强烈的有源干扰使雷达面临着愈加严峻的挑战。因此在复杂多变的电子干扰环境下,使雷达系统在最短时间内识别出所面临的干扰信号类型,进而选择出最有效的抗干扰方法进行对抗,已成为雷达抗干扰技术发展的一个重要方向。本文在现有理论知识和相关论文的基础上,研究了常规的有源欺骗干扰产生机理及作用机理,利用多尺度分解理论,对三类拖引干扰及其混合干扰进行特征级分类识别。论文具体研究内容如下:1.本文通过将一个相干处理周期的脉冲序列提取有效部分拼接成一维向量,弥补单个脉冲信号无法体现拖引干扰产生与实施过程本质特点的不足,讨论了一个相干处理周期中包含的脉冲回波个数对特征区分度及最后识别结果的影响,研究发现,多回波特征级识别结果好于单个回波,且回波数越多分类识别效果越好。2.本文利用多尺度小波分解及经验模态分解两种多尺度分解方法对接收到的三类拖引干扰信号预处理后进行多尺度分解,分别从小波分解得到的小波系数中提取高频细节分量能量比、低频逼近分量归一化能量等小波域特征,从经验模态分解得到的各本征模函数中提取频域矩偏度、矩峰度、噪声因子等频域特征。根据提取的特征在各干噪比下的区分度及波动性选择性能较优的特征作为特征库内特征因子,建立小型特征库,最后利用特征库进行拖引干扰子类型的分类识别。仿真结果表明,利用多尺度分解提取的特征可有效区分三类拖引干扰,且得到令人满意的分类识别结果。3.本文考虑到实际中多种雷达有源欺骗干扰同时存在的情况,针对七种情况的混合信号,利用多尺度分解算法分解接收信号,提取小波系数特征及频域特征,计算各特征在所有干噪比下的均值,建立小型混合干扰特征库,并对干扰信号进行分类识别。仿真实验结果表明,多尺度分解特征库辅助对混合干扰信号识别的方法可较为有效的区分七种混合信号情况,此种方法是有效且具有实际意义的。
[Abstract]:With the development of electronic jamming technology, more intense active jamming makes radar face more severe challenges. Therefore, in the complex and changeable electronic jamming environment, the radar system can identify the type of jamming signal in the shortest time, and then choose the most effective anti-jamming method to counteract it. It has become an important direction in the development of radar anti-jamming technology. On the basis of existing theoretical knowledge and related papers, this paper studies the mechanism of generation and action of conventional active deception jamming, and uses multi-scale decomposition theory to classify and identify three kinds of towed jamming and their mixed interference. The specific contents of this paper are as follows: 1: 1. In this paper, a coherent processing period pulse sequence is segmented into a one-dimensional vector, which makes up for the fact that a single pulse signal is unable to reflect the essential characteristics of the process of generating and implementing the towing interference. The effect of the number of pulse echoes contained in a coherent processing period on the classification and the final recognition results is discussed. It is found that the recognition result of multi-echo feature level is better than that of single echo, and the more the echo number, the better the recognition effect is. 2. In this paper, two multiscale decomposition methods, multiscale wavelet decomposition and empirical mode decomposition, are used to preprocess the received three kinds of towed interference signals. The wavelet domain features such as the energy ratio of the high-frequency detail component and the normalized energy of the low-frequency approximation component are extracted from the wavelet coefficients obtained from the wavelet decomposition, and the moment skewness and kurtosis in the frequency domain are extracted from the eigenmode functions obtained by the empirical mode decomposition. The feature of noise factor in equal frequency domain. According to the distinguishing degree and volatility of the extracted features under each dry noise ratio, the feature with better performance is selected as the feature factor within the feature bank, and a small feature library is established. Finally, the feature library is used to classify and identify the types of towing interference. The simulation results show that the features extracted by multi-scale decomposition can effectively distinguish three types of towing interference and obtain satisfactory classification and recognition results. In this paper, considering the simultaneous existence of many kinds of radar active deceptive jamming in practice, the multi-scale decomposition algorithm is used to decompose the received signals and extract the wavelet coefficients and frequency domain features for the mixed signals of seven kinds of cases. The mean value of each feature under all the dry noise ratio is calculated, and the small mixed interference feature database is established, and the interference signal is classified and identified. The simulation results show that the multi-scale decomposition feature library can effectively distinguish seven kinds of mixed signals, and this method is effective and has practical significance.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN974

【相似文献】