采用压缩感知的阵列失效单元诊断方法

针对现有阵列单元故障诊断方法随阵元数目增多而存在的采样数量大、诊断时间长、计算复杂度高等缺陷，提出了一种采用压缩感知理论的故障诊断方法．该方法基于故障单元数目固有的稀疏性，利用完好阵列和实际阵列激励的差值构造稀疏信号．根据目标方位信息设计测量矩阵的网格划分准则，并通过测量矩阵以随机欠采样方式获取少量测量数据．结合平行坐标下降算法对该稀疏信号进行精确重构，从而实现故障单元的准确诊断．理论分析和仿真实验表明，文中提出的方法不仅明显减少了采样数量，有效缩短了诊断时间，大幅降低了计算复杂度，而且进一步提高了故障信息的重构精度．

Diagnosis method for defective array elements based on compressive sensing

As the number of elements of the array increases, huge sampling data, long measurement time and high computation costs are several features of existing array fault diagnosis methods. Therefore, a diagnosis approach based on Compressive Sensing is investigated in this paper. The proposed method utilizes the sparsity of the number of failed elements and the sparse signal derives from difference incentives of reference array and the array under test. The criterion for grid division of the measurement matrix is designed according to the target direction in the spatial domain and a small number of measurement data are then obtained in the far field radiation pattern via a random under-sampling strategy. The Parallel Coordinate Decent Algorithm is used to implement fault diagnosis by reconstructing this sparse signal. Theoretical analysis as well as simulation results indicate that the proposed method not only reduces the amount of spatial sampling data, truncates the diagnosis time and abates the computational complexity significantly , but also improves the accuracy of recovered information on defective elements.