一种基于复值盲分离的欺骗干扰抑制算法  

An Algorithm of Deception Jamming Suppression Based on Complex-Value Blind Source Separation

作　　者：尹洪伟[1] 李国林[1] 路翠华[1] 

机构地区：[1]海军航空工程学院,山东烟台264001

出　　处：《上海交通大学学报》2015年第10期1564-1569,共6页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金项目(61102165)资助

摘　　要：针对线性调频引信欺骗干扰问题,提出了一种基于复值盲分离的欺骗干扰抑制算法.其中:将阵元间的延时转化为信号混合系数,避免了实域盲分离中阵元间需要精确延时补偿的弊端,根据含噪声观测信号协方差矩阵为非单位矩阵的特点,对复值非高斯最大(CMN)算法及其非线性函数进行修正,以提高算法的抗噪声能力,并借助于分离回波与干扰匹配滤波信号的幅值统计方差鉴别回波与干扰信号.仿真结果表明:在低信噪比条件下,改进的CMN算法明显优于原算法;当信噪比大于10dB时,分离回波信号的相似度在80%以上,而且即使信噪比很低,信号的鉴别率也可达到100%.An algorithm of deception jamming suppression based on the complex-value blind source separation（BSS）was proposed to solve the problem of linear frequency modulation（LFM）fuze deception jamming.The delay-times between the array elements were transformed into the matrix mixing coefficients,so that the disadvantage of the accurate delay-time compensation in real domain BSS was avoided.The complex maximization of the non-Gaussianity（CMN）algorithm and its nonlinear function were amended according to the characteristic that the covariance matrix of the noisy observed signals was not an identity matrix.Therefore,the anti-noise ability of the algorithm was improved.Besides,the echo and jamming were distinguished using the variance characteristics of their matching filter signal amplitude statistics.Simulation results show that the improved CMN algorithm is superior to the original one under the low SNR,when SNR10dB,the similarity coefficient of target echo is larger than 80%,and the discrimination probability of the signals can be 100% even when the SNR is very low.

关 键 词：复值盲分离 欺骗干扰抑制 阵列天线 SMSP干扰 

分 类 号：TN973[电子电信—信号与信息处理]