论文部分内容阅读
针对低信噪比(SNR)条件下传统辐射源识别算法性能下降的问题,提出了基于压缩协作表示的识别算法,分别从特征提取和分类器设计两方面进行描述。首先将时域辐射源信号变换到二维时频域,通过图像处理方法提取高维特征列向量。经随机矩阵压缩到一定维度后,输入到提出的压缩协作表示分类器中得到识别结果。进而,对协作表示系数进行非负约束,提出了更符合实际应用场景的算法。仿真结果验证了所提算法的可行性与有效性,且在低信噪比条件下稳健性强、抗噪声干扰性能好、计算量较小、易于工程实现。
Aiming at the problem of degrading the performance of traditional radiation source recognition algorithms under low signal-to-noise ratio (SNR), an identification algorithm based on compressed collaborative representation is proposed, which is described from two aspects of feature extraction and classifier design respectively. Firstly, the signal of time-domain radiation source is transformed into two-dimensional time-frequency domain, and the high dimensional feature vector is extracted by image processing method. After the random matrix is compressed to a certain dimension, it is input into the proposed compression collaborative classifier to obtain the recognition result. Furthermore, non-negative constraints on the cooperation coefficients are proposed, and an algorithm that is more suitable for practical applications is proposed. The simulation results verify the feasibility and validity of the proposed algorithm, and are robust under low signal-to-noise ratio (SNR), good anti-noise interference performance, small computational complexity and easy implementation.