阵列通道幅相校准是高频地波雷达方位估计必不可少的环节。文中提出一种基于自动识别系统 (AIS)和Canopy+Kmeans 的聚类算法对阵列的幅相误差进行校准。AIS 直接用于阵列幅相校准将会出现许多虚假和错误的校准值信息,因此还需要对AIS 得到的校准值进行进一步筛选。该方法结合机器学习中的Canopy 算法和Kmeans 算法,利用AIS 船只信号得到的幅度和相位校准值进行自动聚类,从而得到正确的幅度和相位校准值。校准之后的雷达数据用多重信号分类算法进行到达角(DOA)估计,DOA 估计的准确度有了大幅的提高。

Array channel amplitude and phase calibration is an essential part of azimuth estimation for high frequency ground wave radar. A Canopy+Kmeans clustering algorithm based on automatic identification system (AIS) information to calibrate the amplitude and phase errors of the array is proposed in this paper. When AIS is directly used for array amplitude and phase calibration, there will be a lot of false and erroneous calibration value information, so it is necessary to further screen the calibration value obtained by AIS. This method combines the Canopy algorithm and the Kmeans algorithm in machine learning to automatically cluster the amplitude and phase calibration values obtained by the AIS ship signal, so as to obtain the correct amplitude and phase calibration values. After the correction, the direction of arrival (DOA) estimation is performed with the multiple signal classification algorithm. The accuracy of the DOA estimation has been greatly improved.

TN955