基于图形轮廓的雷达信号模糊函数主脊切面特征提取方法

A Novel Feature Extraction Method for the Slice of Ambiguity Function Main Ridge of Radar Signals Based on Graphic Contours

从复杂体制雷达辐射源信号中提取有效的特征参数,是解决复杂雷达信号分选问题的有效方法。鉴于模糊函数(AF)的主脊切面能很好地描述信号的模糊能量分布信息,提出一种基于图形轮廓的AF主脊切面特征提取方法,该方法提取了反映切面图形轮廓变化剧烈程度的累积角度和反映切面图形轮廓固有特征的单位累积长度作为信号分选的特征向量。模糊C均值聚类实验结果表明:当信噪比不低于12 dB时,常规信号、线性调频信号、二相编码、四相编码、M伪随机序列和二频编码6种典型雷达信号的分选成功率均为100%,即使在信噪比为0 dB时,平均分选成功率也能达到93. 5%;在0～20 dB动态信噪比环境中平均分选成功率达到98. 9%.理论分析与分选耗时比较实验结果表明,所提方法具有较低的算法复杂度,证实了其有效性和时效性。

Extracting eft）ctive characteristic parameters from the complex radar emitter signals is an effective method to solve the problem of complex radar signal sorting. Because the slice of ambiguity function main ridge of signal can be used to descript the ambiguity energy distribution information of signal, a fea- ture extraction method of the ambiguity function main ridge slice based on graphics contours is proposed. The proposed method can extract two novel features, i. e., cumulated angle and unit cumulative length, which describe the intensity of change and the inherent characteristics of graphics contour, respectively. The experimental resuhs of fuzzy C-means clustering show that the sorting success rate of six types of radar signals, such as nomlal signal, chi@ signal, binaU phase shift keying, quaternaU phase shift keying, M pseudo-random series, and binaty frequency shift keying, is 100% when signal-to-noise ratio （SNR） is no less than 12 dB; and the average success rate of signal sorting is up to 93.5% even if SNR is as low as 0 dB. When SNR changes from 0 dB to 20 dB, the average success rate of signal sorting can reach to 98.9%. The theoretical analysis and the comparative experiment of time consumption for signal sorting show that the proposed method has lower complexity, effectiveness and timeliness.