Optimization of Subarray Partition for Large Planar Phased Array Radar Based on Weighted K-Means Clustering Method

Abstract

Subarray partition is indispensable in large phased array radar system for reducing the manufacturing cost as well as realizing the system potentiality. The optimization of subarray partition for large planar phased array radar according to weighted K-means clustering method is mainly investigated in this paper. Based on the excitation matching technique, the optimization of subarray partition in monopulse application can be reformulated as a clustering of reference gain ratios to minimize the excitation matching error. However, when the element weights are non-uniform for specific intentions such as low sidelobes, the matching error could not be minimized completely by traditional K-means clustering. Therefore, in this paper, a weighted K-means clustering method is proposed to reduce the matching error by modifying the membership rule and cluster center of K-means clustering. The proposed method can provide smaller matching error compared with conventional clustering methods, especially when the elements are weighted non-uniformly. The effectiveness of proposed method is validated by numerical simulations and compared with several classical clustering methods.