基于多特征联合处理的灵巧噪声干扰识别

Multi-Feature-Based Identification of Smart Noise Jamming

灵巧噪声干扰已成为一种作用于新体制相参雷达的重要干扰类型。为了有效抑制该类干扰,提出一种基于多维特征的抗干扰方式。通过分析DRFM产生灵巧噪声干扰原理,建立目标与干扰信号模型;在分析对比两类信号特性的基础上,提取包络起伏参数、相位门限内概率及盒维数特征以表征目标与干扰信号在波形、相位及尺度信息上的差异;为了进一步提高干扰识别性能,加入表征信号复杂度的近似熵特征,分析表明该特征因子对噪声具有较强的鲁棒性;最后采用支撑向量机对提取的多维特征进行处理以实现灵巧噪声干扰的识别对抗。仿真实验表明,该方法对目标和干扰的正确识别率高且基本不受干噪比影响。

Smart noise jamming has become an important jamming to some new radar systems.To deal with the jamming,an anti-jamming method based on multi-feature is proposed in this paper.We build the models of target echo and jamming by analyzing the principle that digital radio frequency memory（DRFM）produces jamming.The envelop fluctuation parameters,the probability in the phase gate,and the box dimension are extracted to represent the difference of waveform,phase and scale between the target and the jamming.Besides,we extract the approximate entropy feature which means signal complexity and has strong robustness against noise.Finally,the support vector machine（SVM） is designed to classify and recognize the different modes.Experiments show that this model has high correct recognition rate for target echo and jamming and it is little affected by jamming to noise rate（JNR）.