Robust Adaptive Beamforming Using k-means Clustering: A Solution to High Complexity of the Reconstruction-Based Algorithm

Abstract

Recently, a new robust adaptive beamforming (RAB) algorithm has been proposed to reconstruct the interference-plus-noise covariance matrix (IPNCM) based on narrowing the interference angular domain and using an annular uncertainty set (NIAD-AUS). The method is robust against unknown arbitrary-type mismatches. However, its computational complexity will increase exponentially with the number of array sensors. In this paper, a novel method is proposed to solve this problem. First, k-means clustering (KMC) algorithm is utilized to estimate the annulus uncertainty set with fewer clustering weight points rather than whole sampling. Second, the KMC Capon spectrum is used to reconstruct the IPNCM. Compared with the previous reconstruction-based algorithms, the proposed approach can retain the high performance of the state-of-the-art NIAD-AUS algorithm. More importantly, it can also obtain the IPNCM more quickly. Lastly, simulation results demonstrate the effectiveness and robustness of the proposed algorithm.