统计模式识别在相控阵天线故障诊断中的应用研究

相控阵天线已广泛应用于雷达系统，而阵列单元的快速诊断日益成为难题。针对相控阵天线阵元故障难以检测的问题，提出了一种基于统计模式识别的方法。首先阐述了相控阵诊断原理，用矩量法构建了仿真环境，并提取了时域特征和小波特征。为增大类间平均距离，建立了故障树诊断模型以减小判别问题的规模，在时域特征空间中用投影聚类算法划分了子空间，在叶节点处用小波特征进行判别，实现了故障阵元的定位。仿真实验表明，该方法在信噪比较低时，比非层次方法优势明显，信噪比大于8时，识别率达到95%以上，且随着规模的增加，识别率并未明显下降，证明该方法理论上能够有效诊断相控阵阵元故障。实际应用中，只须对阵列的行或列逐次诊断即可获知整个阵面的故障信息。 

Research on Application of Statistical Pattern Recognition in Phased Array Antenna Fault Diagnosis

As phased array antenna is widely used in radar system, fast diagnosis of element has become a challenge. According to the difficulty of phased array antenna elements detection, a novel method based on statistical pattern recognition is proposed. At first, the phased array diagnosis principle is stated. Simulation environment is built with MoM. Time domain feature and wavelet feature are extracted. In order to enlarge the mean distance between classes, the fault tree diagnosis model is built to reduce the scale of the discrimination problem, subspaces are divided in time domain feature space using projected clustering algorithm, and wavelet feature is utilized to discriminate in leaf nodes. The location of the faulty elements is realized. Simulations show that at low SNR, the method has obvious advantages over nonhierarchic method. At SNR higher than 8dB, recognition rate reaches 95%. As the scale increases, recognition rate does not decrease obviously. Results turn out that this method is effective in diagnosis of phased array element theoretically. In application, only conduct the diagnosis on each row or column of the array, the failure information of the whole antenna can be learned.