基于双谱综合特征提取的距离-速度同步拖引干扰识别方法研究

Identification of Range-Velocity Gate Pull Off Jamming Based on Integrated Feature Extraction of Bispectrum

为有效对抗距离-速度同步拖引(R-VGPO)欺骗干扰,提出了一种综合的双谱特征提取方法用于识别真实目标回波信号和欺骗干扰。建立了拖引干扰信号模型,基于双谱分析良好的抗噪性能,分别对真实回波信号和欺骗干扰作双谱变换。根据双谱估计的结果,用围线积分双谱与辅对角积分双谱提取特征作为一次特征向量;为降低特征维数并消除交叉项干扰,用信息熵和基于Fisher准则的特征选择方法进行二次特征提取,建立特征数适中的综合特征参数集,在保留最具区分度的特征向量的同时,避免信息损失,突出优秀特征值的作用。仿真结果表明:与基于特征选择和基于熵的单一特征提取方法相比,提出的双谱综合特征提取方法可获得更高的识别准确率和更好的抗噪性能,能有效对抗距离-速度同步拖引欺骗干扰。

In order to counteract range-velocity gate pull off deception jamming effectively,the integrated bispectrum extraction method was put forward to identify real echo and deception jamming.The pull off jamming model was established.Based on the good anti-noise performance of bispectrum estimation,the initial eigenvector was extracted by surrounding-line integral bispectrum and secondary diagonally integral bispectrum.To reduce feature dimension and eliminate the cross disturbance,the second feature was extracted by information entropy and feature selection based on Fisher criterion.The integrated feature set with suitable feature number was obtained,which could avoid the information loss while keeping the feature vector with the most distinguish degree.So the excellence features were emphasized.The simulation results show that the proposed integrated bispectrum feature extraction method can obtain higher identification accuracy and counteract range-velocity gate pull off deceptive jamming effectively compared to a single feature extraction based on feature selection or entropy.