基于稀疏时频分解的空中目标微动特征分析

本文以直升机的RCS 信号为研究对象，利用Gabor 原子构造过完备时频原子库，在该库上实现基于匹配追踪（Matching Pursuit，MP）的信号分解，并结合遗传算法寻找最佳匹配原子。最后将每次分解得到的Gabor 原子进行Wigner-Ville 变换，利用结果叠加得到无交叉干扰项的信号的Wigner-Ville 分布（Wigner-VilleDistribution，WVD）。仿真结果表明，该方法能够提高稀疏分解的效率，并且用少量的Gabor 原子就能表示信号WVD。与直接进行Wigner-Ville 变换相比，本方法可以有效地抑制交叉干扰项，且保持高时频分辨率。

Micro-motion Features Analysis for Air Targets based on Sparse Time-frequency Decomposition

Take the helicopter RCS signal as an example, the sparse time-frequency decomposition method is researched in this paper. Firstly, an over-complete dictionary is constructed with Gabor atoms. Then the RCS signal is decomposed using matching pursuit algorithm based on Gabor atom dictionary, and the best matching atoms are found by genetic algorithm. At last, the signal’s WVD distribution with no cross-interference is obtained by superposing the results of each decomposed Gabor atom’s WVD distribution. The simulation results show that the proposed method can improve the efficiency of sparse decomposition and express the signal WVD with a few Gabor atoms. Compared with Wigner-Ville distribution, this method can maintain high time-frequency resolution with no cross-interference.