基于稀疏表示分类的雷达欺骗干扰识别方法

Deception jamming recognition of radar based on sparse representation classification

为了应对基于数字射频存储的各种欺骗干扰信号,提出了一种基于稀疏表示分类的欺骗干扰识别算法。通过小波包分解重构把信号划分为不同频段,然后对信号提取三阶累积量切片特征构造特征矩阵,并利用奇异值分解对特征进行降维,提取主要分量。最后利用稀疏表示分类在不同频段上对信号进行分类识别,利用决策融合的方法对分类结果进行整合。经验证,该方法具有很好的抗噪性能,能够有效识别几种常见的欺骗干扰信号,在信噪比为0 dB时,欺骗干扰平均识别率达到90%以上,并与其他欺骗干扰识别方法进行了对比,显示了所提方法的优越性。

In order to deal with a variety of deception jamming signals based on digital radio frequency memory jammer,this paper presents an algorithm for deception jamming recognition based on sparse representation classification.Firstly,the wavelet packet decomposition and reconstruction are used to divide the signal into different frequency bands,and the feature matrix is constructed by features extracted from the third-order cumulant slice.Then,singular value decomposition is used to reduce the dimension of the feature matrix and extracts the main components.Next,the classification results on each frequency band are obtained by sparse representation classification method,and finally the results are integrated by decision fusion.It has been verified that this method has good anti-noise performance and can effectively identify several common deception jamming signals.When the signal noise ratio is 0 dB,the average recognition rate of deception jamming is more than 90%.Compared with other methods of deception jamming recognition,the superiority of proposed method can be proved.