基于脉内特征的雷达信号分选方法

当前复杂的电子环境下不同调制的多种雷达信号，具有交叠严重、分选困难的特点。针对这一问题提出一种基于时频原子特征的雷达信号分选方法。该方法根据欧氏距离准则，在超完备的Chriplet时频原子库中，利用杂草算法提取可分性较好的原子，并与各信号做内积作为雷达信号分选的特征值，利用改进的FCM聚类算法进行聚类。在不同信噪比下，对5种不同调制参数的雷达信号进行分选，仿真结果表明，在较低的信噪比下，该方法能够进行有效的雷达信号分选。

Radar signal sorting method based on the intra-pulse characteristics

In the complicated electronic environment, radar signals with different modulation overlap seriously, and they are difficult to sort. This paper proposes a radar signal sorting method based on the features of the time？fre？quency atoms. According to the Euclidean distance criterion, this method extracts the atoms with better separability using the weed algorithm in super complete library of the Chriplet time？frequency atoms, then combines these atoms with each signal to make the inner product, deriving the eigenvalues of radar signal sorting, and next clusters by the modified FCM clustering algorithm. Under different SNRs, this paper makes a sort for five different radar signals with different modulation parameters. The simulation results show that in low SNR, this method is effective.