Adaptive Beampattern Synthesis for Partially-Clustered Array Through $ K$-Means and Amplitude-Only MVDR

Partially-clustered arrays (PCA) are recognized for offering a balanced compromise between beamforming performance and implementation complexity. In this letter, leveraging the information of both pointing and nulling directions, we present a low-complexity method aimed at optimizing not only the phase-shift and amplitude values but also the configuration of clustering for adaptive beampattern synthesis. Simulation results underscore the efficacy of our approach, demonstrating its output signal-to-interference-plus-noise ratio (SINR) performance consistently close to that of the minimum variance distortionless response (MVDR) beamformer, and markedly superior to the conventional excitation matching (EM) based PCA approach.