压缩感知算法在阵列测向中的相干信号分辨

现有算法在进行信号重建时，需要原始信号的稀疏信息来控制初始原子集的大小和算法迭代的最大次数，这削弱了重建精度，增加了计算复杂度，限制了其实际应用能力。为了克服这个问题，提出了基于正交匹配追踪算法的改进算法。该算法将阵元个数作为假设稀疏度，在稀疏度未知的初始阶段扩展初始原子集。最后通过对测量信号的幅度进行阈值过滤以实现信号的分辨和信源数的估计。为解决测向人员面对监测软件所指示的多个信号示向度中，无法分辨哪些指示方向是同一个信号源所产生的折反射信号的问题，首次提出了基于压缩感知算法的相干关系分析和相干信号分辨的模型。实测数据结果表明，该算法在不同环境下可以清晰分辨相干信号，并且能够适应多种实验平台。

Coherent signal resolution in array direction finding using compressed sensing algorithm

For the existing algorithms control the size of the initial atom set and the maximum number of algorithm iterations by spare information, which would inevitably reduce the accuracy of reconstruction, increase the computational complexity, and limit its practical application. To overcome this problem, this paper proposed an improved algorithm based on orthogonal matching tracking algorithm. The proposed method regarded the number of elements as assumed sparsity as first. And then, it expanded the initial atom at initial stage that the sparsity information was unknown. Finally, it filtered the amplitude of the measured signal by a threshold to realize signal resolution and source number estimation. Based on the compressed sensing algorithm, it originally introduced a coherent relation analysis and coherent signal resolution model, which could help the direction-finder to effectively distinguish the directions of the reflected signals generated by the same signal source among the multiple signal directions indicated by the monitoring software. The experimental results show that the algorithm can clearly distinguish the coherent signals in different environments, and can be adapted to a variety of experimental platforms.