针对低信噪比下雷达辐射源信号分类,首先提出了基于高阶累积量和小波包变换相结合的特征提取方法,然后设计支持向量机分类器,并运用粒子群优化算法对分类器的参数进行寻优,最终实现对雷达辐射源信号的自动分类。仿真实验结果表明,在信噪比为-4dB时,6种雷达辐射源信号的平均识别率仍能达到93.83%,在低信噪比环境下取得了较为理想的分类效果。%To correctly classify advanced radar emitter signals in the condition of low signal noise ratio,a novel method using high order cumulant and wavelet packet transform is proposed for feature extraction,then a Support Vector Machines classifier is designed,parameters of which are optimized using particle swarm optimization for better classification result,and identification of radar emitter signals is realized automatically.Experiments conducted on six typical emitter signals show that the proposed method works effectively as high as 93.83% recognition rate when SNR=-4dB,which proves classification results are better in low signal noise ratio.
展开▼
