基于级联原子库的时频原子分解

Time-frequent atom decomposition based on cascade dictionary

为了分解多分量雷达辐射源信号,提出一种基于级联原子库的时频原子分解方法。该方法先构建级联原子库,采用该原子库对信号进行时频原子分解,以得到信号的最优稀疏表示。仿真实验结果表明,对多分量雷达辐射源信号进行时频原子分解,该方法在重构精度、衰减率和重构信号时频聚集性3个方面均优于单一时频原子库下的时频原子分解方法。

In order to decompose muhi-component radar emitter signal, the paper presents a time frequent atom decomposi- tion algorithm based on cascade dictionary. The introduced method first builds cascade dictionary, and then decomposes signal by using the cascade dictionary, to obtain the optimal sparse representation of signal. Experimental results conducted on multi-component radar emitter signals show that the introduced method is superior to single dictionary based approach, in term of the reconstruction accuracy, decaying rate and time-frequency energy distribution.