经验模态分解模态域LCMV研究报告.docVIP

基于经验模态分解的模态域LCMV研究 杨明 周顺 （总参谋部第63研究所 江苏 南京 210007） 摘 要：已有研究对宽带阵列信号的子带分解处理中子带划分是固定的，各子带内信号数分布不均衡。为此文中研究了子带自适应划分的模态域宽带自适应阵列算法，提出将经验模态分解（EMD）融合于线性约束最小方差（LCMV）算法中，EMD把多分量信号自适应地分解成多阶内蕴模态函数（IMF），突出了信号的局部瞬间特征，两者结合既满足了LCMV的窄带信号要求又充分发挥了EMD自适应的特点，解决了阵元数目受到严格限制时，宽带阵列处理自由度不足、抗干扰性能下降的问题，在强干扰环境和各信号强度相差较大时也能有效检测出较弱信号。与基于FFT的子带分割相比，无需预先划分模态函数的频段，阵元数目受限时增加了处理自由度，仿真表明3阵元强干扰条件下改善增益11dB，正确有效得到验证。 关键词： 中图分类号：TP 391 文献标识码 A 文章编号： A Study of Mode Domain LCMV Algorithm Based on Empirical Mode Decomposition Yangming Zhoushun (The 63rd research Institute of PLA General Staff Headquarters, NanJing 210007) Abstract: Sub-band array processing for wideband adaptive array is proposed in references, but the sub-band segments is fixed so the signals number distribution in each segment is unbalanced, the empirical mode decomposition(EMD) domain Linearly Constrained Minimum Variance(LCMV) algorithm for wideband array is proposed based on adaptive band segmentation, the EMD decompose multicomponent signals into many intrinsic mode function(IMFs) and focus on signals local characteristic, Combining IMF characteristic and LCMV could meet narrow-band request as well as take fully the advantage of adaptation. the problems of anti-jamming performance degraded and insufficient processing freedom degrees is settled for wideband adaptive array in the limited elements situation, the weaker target’s detection is realized effectively in strong interference. Compared with the traditional FFT sub-band, the proposed is applicable to wideband limited elements adaptive array with no prior frequency partition, increasing the array processing freedom and improving gain performance by 11dB at 3 elements and in strong interference situation. Simulations are conducted to indicate the validity. Key Words: Mode Domain LCMV; empirical mode decomposition; wideband adaptive array  自适应阵列在期望信号的方向形成主波束，在干扰方向上形成零陷，解决了干扰信号和期望信号频率相同或者多个相同频带的信号叠加在一起时，时域滤波器无法提取期望信号的问题。但是宽带自适应阵列在增加处理带宽的同时也增加阵列的散布度，而且抑制一个宽带干扰所需的处