【摘要】 为提高雷达辐射源个体在不同信噪比下的识别率,提出了基于自适应变分模态分解(VMD)和人工蜂群优化的分类器(ABC-SVM)的方法,利用蛙跳算法对最佳影响参数组合进行搜索,根据得到的最优参数,通过VMD变换对雷达信号进行处理,由于雷达信号分解成各个模态的中心频率不同,选择中心频率作为辐射源信号的无意调制特征参数送入人工蜂群算法优化的分类器进行学习,实现对不同雷达辐射源的分类.实验结果表明,信号分解后的各模态分量的中心频率在一定范围内稳定,ABC-SVM也有着更好的识别效果,当信噪比在15 dB时,三个辐射源在两种信号形式下的识别率全部达到90%,并通过与其他算法进行对比,在不同的信噪比下,识别率均有提高,所提出的方法可以对雷达辐射源个体进行有效识别.更多还原

【Abstract】 In order to improve the recognition rate of radar emitters under different signal-tonoise ratio ratio(SSN),a method based on adaptive variational mode decomposition and artificial bee colony optimization classifier(ABC-SVM) was proposed. Firstly,the frog hopping algorithm is used to search for the best influence parameter combination. Excellent parameters,the radar signal was processed by VMD transformation. Since the radar signal was decomposed into different modal frequencies,the unintentional modulation characteristic parameter of the center frequency was selected as the source code of the artificial bee colony algorithm. Learn to achieve classification of different radar emitters individual. The experimental results showed that the center frequency of each modal component after signal decomposition was stable within a certain range,and ABC-SVM also has better recognition effect. When the signal-to-noise ratio was 15 dB,three radiation sources were in two signal forms. The recognition rate was up to 90%,and compared with other algorithms,the recognition rate was improved under different signal-to-noise ratios. The proposed method can effectively identify the radar emitters individual.更多还原